1.. Copyright 2020 DisplayLink (UK) Ltd. 2 3=================== 4Userland interfaces 5=================== 6 7The DRM core exports several interfaces to applications, generally 8intended to be used through corresponding libdrm wrapper functions. In 9addition, drivers export device-specific interfaces for use by userspace 10drivers & device-aware applications through ioctls and sysfs files. 11 12External interfaces include: memory mapping, context management, DMA 13operations, AGP management, vblank control, fence management, memory 14management, and output management. 15 16Cover generic ioctls and sysfs layout here. We only need high-level 17info, since man pages should cover the rest. 18 19libdrm Device Lookup 20==================== 21 22.. kernel-doc:: drivers/gpu/drm/drm_ioctl.c 23 :doc: getunique and setversion story 24 25 26.. _drm_primary_node: 27 28Primary Nodes, DRM Master and Authentication 29============================================ 30 31.. kernel-doc:: drivers/gpu/drm/drm_auth.c 32 :doc: master and authentication 33 34.. kernel-doc:: drivers/gpu/drm/drm_auth.c 35 :export: 36 37.. kernel-doc:: include/drm/drm_auth.h 38 :internal: 39 40 41.. _drm_leasing: 42 43DRM Display Resource Leasing 44============================ 45 46.. kernel-doc:: drivers/gpu/drm/drm_lease.c 47 :doc: drm leasing 48 49Open-Source Userspace Requirements 50================================== 51 52The DRM subsystem has stricter requirements than most other kernel subsystems on 53what the userspace side for new uAPI needs to look like. This section here 54explains what exactly those requirements are, and why they exist. 55 56The short summary is that any addition of DRM uAPI requires corresponding 57open-sourced userspace patches, and those patches must be reviewed and ready for 58merging into a suitable and canonical upstream project. 59 60GFX devices (both display and render/GPU side) are really complex bits of 61hardware, with userspace and kernel by necessity having to work together really 62closely. The interfaces, for rendering and modesetting, must be extremely wide 63and flexible, and therefore it is almost always impossible to precisely define 64them for every possible corner case. This in turn makes it really practically 65infeasible to differentiate between behaviour that's required by userspace, and 66which must not be changed to avoid regressions, and behaviour which is only an 67accidental artifact of the current implementation. 68 69Without access to the full source code of all userspace users that means it 70becomes impossible to change the implementation details, since userspace could 71depend upon the accidental behaviour of the current implementation in minute 72details. And debugging such regressions without access to source code is pretty 73much impossible. As a consequence this means: 74 75- The Linux kernel's "no regression" policy holds in practice only for 76 open-source userspace of the DRM subsystem. DRM developers are perfectly fine 77 if closed-source blob drivers in userspace use the same uAPI as the open 78 drivers, but they must do so in the exact same way as the open drivers. 79 Creative (ab)use of the interfaces will, and in the past routinely has, lead 80 to breakage. 81 82- Any new userspace interface must have an open-source implementation as 83 demonstration vehicle. 84 85The other reason for requiring open-source userspace is uAPI review. Since the 86kernel and userspace parts of a GFX stack must work together so closely, code 87review can only assess whether a new interface achieves its goals by looking at 88both sides. Making sure that the interface indeed covers the use-case fully 89leads to a few additional requirements: 90 91- The open-source userspace must not be a toy/test application, but the real 92 thing. Specifically it needs to handle all the usual error and corner cases. 93 These are often the places where new uAPI falls apart and hence essential to 94 assess the fitness of a proposed interface. 95 96- The userspace side must be fully reviewed and tested to the standards of that 97 userspace project. For e.g. mesa this means piglit testcases and review on the 98 mailing list. This is again to ensure that the new interface actually gets the 99 job done. The userspace-side reviewer should also provide an Acked-by on the 100 kernel uAPI patch indicating that they believe the proposed uAPI is sound and 101 sufficiently documented and validated for userspace's consumption. 102 103- The userspace patches must be against the canonical upstream, not some vendor 104 fork. This is to make sure that no one cheats on the review and testing 105 requirements by doing a quick fork. 106 107- The kernel patch can only be merged after all the above requirements are met, 108 but it **must** be merged to either drm-next or drm-misc-next **before** the 109 userspace patches land. uAPI always flows from the kernel, doing things the 110 other way round risks divergence of the uAPI definitions and header files. 111 112These are fairly steep requirements, but have grown out from years of shared 113pain and experience with uAPI added hastily, and almost always regretted about 114just as fast. GFX devices change really fast, requiring a paradigm shift and 115entire new set of uAPI interfaces every few years at least. Together with the 116Linux kernel's guarantee to keep existing userspace running for 10+ years this 117is already rather painful for the DRM subsystem, with multiple different uAPIs 118for the same thing co-existing. If we add a few more complete mistakes into the 119mix every year it would be entirely unmanageable. 120 121.. _drm_render_node: 122 123Render nodes 124============ 125 126DRM core provides multiple character-devices for user-space to use. 127Depending on which device is opened, user-space can perform a different 128set of operations (mainly ioctls). The primary node is always created 129and called card<num>. Additionally, a currently unused control node, 130called controlD<num> is also created. The primary node provides all 131legacy operations and historically was the only interface used by 132userspace. With KMS, the control node was introduced. However, the 133planned KMS control interface has never been written and so the control 134node stays unused to date. 135 136With the increased use of offscreen renderers and GPGPU applications, 137clients no longer require running compositors or graphics servers to 138make use of a GPU. But the DRM API required unprivileged clients to 139authenticate to a DRM-Master prior to getting GPU access. To avoid this 140step and to grant clients GPU access without authenticating, render 141nodes were introduced. Render nodes solely serve render clients, that 142is, no modesetting or privileged ioctls can be issued on render nodes. 143Only non-global rendering commands are allowed. If a driver supports 144render nodes, it must advertise it via the DRIVER_RENDER DRM driver 145capability. If not supported, the primary node must be used for render 146clients together with the legacy drmAuth authentication procedure. 147 148If a driver advertises render node support, DRM core will create a 149separate render node called renderD<num>. There will be one render node 150per device. No ioctls except PRIME-related ioctls will be allowed on 151this node. Especially GEM_OPEN will be explicitly prohibited. Render 152nodes are designed to avoid the buffer-leaks, which occur if clients 153guess the flink names or mmap offsets on the legacy interface. 154Additionally to this basic interface, drivers must mark their 155driver-dependent render-only ioctls as DRM_RENDER_ALLOW so render 156clients can use them. Driver authors must be careful not to allow any 157privileged ioctls on render nodes. 158 159With render nodes, user-space can now control access to the render node 160via basic file-system access-modes. A running graphics server which 161authenticates clients on the privileged primary/legacy node is no longer 162required. Instead, a client can open the render node and is immediately 163granted GPU access. Communication between clients (or servers) is done 164via PRIME. FLINK from render node to legacy node is not supported. New 165clients must not use the insecure FLINK interface. 166 167Besides dropping all modeset/global ioctls, render nodes also drop the 168DRM-Master concept. There is no reason to associate render clients with 169a DRM-Master as they are independent of any graphics server. Besides, 170they must work without any running master, anyway. Drivers must be able 171to run without a master object if they support render nodes. If, on the 172other hand, a driver requires shared state between clients which is 173visible to user-space and accessible beyond open-file boundaries, they 174cannot support render nodes. 175 176Device Hot-Unplug 177================= 178 179.. note:: 180 The following is the plan. Implementation is not there yet 181 (2020 May). 182 183Graphics devices (display and/or render) may be connected via USB (e.g. 184display adapters or docking stations) or Thunderbolt (e.g. eGPU). An end 185user is able to hot-unplug this kind of devices while they are being 186used, and expects that the very least the machine does not crash. Any 187damage from hot-unplugging a DRM device needs to be limited as much as 188possible and userspace must be given the chance to handle it if it wants 189to. Ideally, unplugging a DRM device still lets a desktop continue to 190run, but that is going to need explicit support throughout the whole 191graphics stack: from kernel and userspace drivers, through display 192servers, via window system protocols, and in applications and libraries. 193 194Other scenarios that should lead to the same are: unrecoverable GPU 195crash, PCI device disappearing off the bus, or forced unbind of a driver 196from the physical device. 197 198In other words, from userspace perspective everything needs to keep on 199working more or less, until userspace stops using the disappeared DRM 200device and closes it completely. Userspace will learn of the device 201disappearance from the device removed uevent, ioctls returning ENODEV 202(or driver-specific ioctls returning driver-specific things), or open() 203returning ENXIO. 204 205Only after userspace has closed all relevant DRM device and dmabuf file 206descriptors and removed all mmaps, the DRM driver can tear down its 207instance for the device that no longer exists. If the same physical 208device somehow comes back in the mean time, it shall be a new DRM 209device. 210 211Similar to PIDs, chardev minor numbers are not recycled immediately. A 212new DRM device always picks the next free minor number compared to the 213previous one allocated, and wraps around when minor numbers are 214exhausted. 215 216The goal raises at least the following requirements for the kernel and 217drivers. 218 219Requirements for KMS UAPI 220------------------------- 221 222- KMS connectors must change their status to disconnected. 223 224- Legacy modesets and pageflips, and atomic commits, both real and 225 TEST_ONLY, and any other ioctls either fail with ENODEV or fake 226 success. 227 228- Pending non-blocking KMS operations deliver the DRM events userspace 229 is expecting. This applies also to ioctls that faked success. 230 231- open() on a device node whose underlying device has disappeared will 232 fail with ENXIO. 233 234- Attempting to create a DRM lease on a disappeared DRM device will 235 fail with ENODEV. Existing DRM leases remain and work as listed 236 above. 237 238Requirements for Render and Cross-Device UAPI 239--------------------------------------------- 240 241- All GPU jobs that can no longer run must have their fences 242 force-signalled to avoid inflicting hangs on userspace. 243 The associated error code is ENODEV. 244 245- Some userspace APIs already define what should happen when the device 246 disappears (OpenGL, GL ES: `GL_KHR_robustness`_; `Vulkan`_: 247 VK_ERROR_DEVICE_LOST; etc.). DRM drivers are free to implement this 248 behaviour the way they see best, e.g. returning failures in 249 driver-specific ioctls and handling those in userspace drivers, or 250 rely on uevents, and so on. 251 252- dmabuf which point to memory that has disappeared will either fail to 253 import with ENODEV or continue to be successfully imported if it would 254 have succeeded before the disappearance. See also about memory maps 255 below for already imported dmabufs. 256 257- Attempting to import a dmabuf to a disappeared device will either fail 258 with ENODEV or succeed if it would have succeeded without the 259 disappearance. 260 261- open() on a device node whose underlying device has disappeared will 262 fail with ENXIO. 263 264.. _GL_KHR_robustness: https://www.khronos.org/registry/OpenGL/extensions/KHR/KHR_robustness.txt 265.. _Vulkan: https://www.khronos.org/vulkan/ 266 267Requirements for Memory Maps 268---------------------------- 269 270Memory maps have further requirements that apply to both existing maps 271and maps created after the device has disappeared. If the underlying 272memory disappears, the map is created or modified such that reads and 273writes will still complete successfully but the result is undefined. 274This applies to both userspace mmap()'d memory and memory pointed to by 275dmabuf which might be mapped to other devices (cross-device dmabuf 276imports). 277 278Raising SIGBUS is not an option, because userspace cannot realistically 279handle it. Signal handlers are global, which makes them extremely 280difficult to use correctly from libraries like those that Mesa produces. 281Signal handlers are not composable, you can't have different handlers 282for GPU1 and GPU2 from different vendors, and a third handler for 283mmapped regular files. Threads cause additional pain with signal 284handling as well. 285 286.. _drm_driver_ioctl: 287 288IOCTL Support on Device Nodes 289============================= 290 291.. kernel-doc:: drivers/gpu/drm/drm_ioctl.c 292 :doc: driver specific ioctls 293 294Recommended IOCTL Return Values 295------------------------------- 296 297In theory a driver's IOCTL callback is only allowed to return very few error 298codes. In practice it's good to abuse a few more. This section documents common 299practice within the DRM subsystem: 300 301ENOENT: 302 Strictly this should only be used when a file doesn't exist e.g. when 303 calling the open() syscall. We reuse that to signal any kind of object 304 lookup failure, e.g. for unknown GEM buffer object handles, unknown KMS 305 object handles and similar cases. 306 307ENOSPC: 308 Some drivers use this to differentiate "out of kernel memory" from "out 309 of VRAM". Sometimes also applies to other limited gpu resources used for 310 rendering (e.g. when you have a special limited compression buffer). 311 Sometimes resource allocation/reservation issues in command submission 312 IOCTLs are also signalled through EDEADLK. 313 314 Simply running out of kernel/system memory is signalled through ENOMEM. 315 316EPERM/EACCES: 317 Returned for an operation that is valid, but needs more privileges. 318 E.g. root-only or much more common, DRM master-only operations return 319 this when called by unpriviledged clients. There's no clear 320 difference between EACCES and EPERM. 321 322ENODEV: 323 The device is not present anymore or is not yet fully initialized. 324 325EOPNOTSUPP: 326 Feature (like PRIME, modesetting, GEM) is not supported by the driver. 327 328ENXIO: 329 Remote failure, either a hardware transaction (like i2c), but also used 330 when the exporting driver of a shared dma-buf or fence doesn't support a 331 feature needed. 332 333EINTR: 334 DRM drivers assume that userspace restarts all IOCTLs. Any DRM IOCTL can 335 return EINTR and in such a case should be restarted with the IOCTL 336 parameters left unchanged. 337 338EIO: 339 The GPU died and couldn't be resurrected through a reset. Modesetting 340 hardware failures are signalled through the "link status" connector 341 property. 342 343EINVAL: 344 Catch-all for anything that is an invalid argument combination which 345 cannot work. 346 347IOCTL also use other error codes like ETIME, EFAULT, EBUSY, ENOTTY but their 348usage is in line with the common meanings. The above list tries to just document 349DRM specific patterns. Note that ENOTTY has the slightly unintuitive meaning of 350"this IOCTL does not exist", and is used exactly as such in DRM. 351 352.. kernel-doc:: include/drm/drm_ioctl.h 353 :internal: 354 355.. kernel-doc:: drivers/gpu/drm/drm_ioctl.c 356 :export: 357 358.. kernel-doc:: drivers/gpu/drm/drm_ioc32.c 359 :export: 360 361Testing and validation 362====================== 363 364Testing Requirements for userspace API 365-------------------------------------- 366 367New cross-driver userspace interface extensions, like new IOCTL, new KMS 368properties, new files in sysfs or anything else that constitutes an API change 369should have driver-agnostic testcases in IGT for that feature, if such a test 370can be reasonably made using IGT for the target hardware. 371 372Validating changes with IGT 373--------------------------- 374 375There's a collection of tests that aims to cover the whole functionality of 376DRM drivers and that can be used to check that changes to DRM drivers or the 377core don't regress existing functionality. This test suite is called IGT and 378its code and instructions to build and run can be found in 379https://gitlab.freedesktop.org/drm/igt-gpu-tools/. 380 381Using VKMS to test DRM API 382-------------------------- 383 384VKMS is a software-only model of a KMS driver that is useful for testing 385and for running compositors. VKMS aims to enable a virtual display without 386the need for a hardware display capability. These characteristics made VKMS 387a perfect tool for validating the DRM core behavior and also support the 388compositor developer. VKMS makes it possible to test DRM functions in a 389virtual machine without display, simplifying the validation of some of the 390core changes. 391 392To Validate changes in DRM API with VKMS, start setting the kernel: make 393sure to enable VKMS module; compile the kernel with the VKMS enabled and 394install it in the target machine. VKMS can be run in a Virtual Machine 395(QEMU, virtme or similar). It's recommended the use of KVM with the minimum 396of 1GB of RAM and four cores. 397 398It's possible to run the IGT-tests in a VM in two ways: 399 400 1. Use IGT inside a VM 401 2. Use IGT from the host machine and write the results in a shared directory. 402 403As follow, there is an example of using a VM with a shared directory with 404the host machine to run igt-tests. As an example it's used virtme:: 405 406 $ virtme-run --rwdir /path/for/shared_dir --kdir=path/for/kernel/directory --mods=auto 407 408Run the igt-tests in the guest machine, as example it's ran the 'kms_flip' 409tests:: 410 411 $ /path/for/igt-gpu-tools/scripts/run-tests.sh -p -s -t "kms_flip.*" -v 412 413In this example, instead of build the igt_runner, Piglit is used 414(-p option); it's created html summary of the tests results and it's saved 415in the folder "igt-gpu-tools/results"; it's executed only the igt-tests 416matching the -t option. 417 418Display CRC Support 419------------------- 420 421.. kernel-doc:: drivers/gpu/drm/drm_debugfs_crc.c 422 :doc: CRC ABI 423 424.. kernel-doc:: drivers/gpu/drm/drm_debugfs_crc.c 425 :export: 426 427Debugfs Support 428--------------- 429 430.. kernel-doc:: include/drm/drm_debugfs.h 431 :internal: 432 433.. kernel-doc:: drivers/gpu/drm/drm_debugfs.c 434 :export: 435 436Sysfs Support 437============= 438 439.. kernel-doc:: drivers/gpu/drm/drm_sysfs.c 440 :doc: overview 441 442.. kernel-doc:: drivers/gpu/drm/drm_sysfs.c 443 :export: 444 445 446VBlank event handling 447===================== 448 449The DRM core exposes two vertical blank related ioctls: 450 451DRM_IOCTL_WAIT_VBLANK 452 This takes a struct drm_wait_vblank structure as its argument, and 453 it is used to block or request a signal when a specified vblank 454 event occurs. 455 456DRM_IOCTL_MODESET_CTL 457 This was only used for user-mode-settind drivers around modesetting 458 changes to allow the kernel to update the vblank interrupt after 459 mode setting, since on many devices the vertical blank counter is 460 reset to 0 at some point during modeset. Modern drivers should not 461 call this any more since with kernel mode setting it is a no-op. 462 463Userspace API Structures 464======================== 465 466.. kernel-doc:: include/uapi/drm/drm_mode.h 467 :doc: overview 468 469.. _crtc_index: 470 471CRTC index 472---------- 473 474CRTC's have both an object ID and an index, and they are not the same thing. 475The index is used in cases where a densely packed identifier for a CRTC is 476needed, for instance a bitmask of CRTC's. The member possible_crtcs of struct 477drm_mode_get_plane is an example. 478 479DRM_IOCTL_MODE_GETRESOURCES populates a structure with an array of CRTC ID's, 480and the CRTC index is its position in this array. 481 482.. kernel-doc:: include/uapi/drm/drm.h 483 :internal: 484 485.. kernel-doc:: include/uapi/drm/drm_mode.h 486 :internal: 487