drm/xe/xe_vm_doc.h

16  * bind engine, and return a handle to the user.
25  * VM bind (create GPU mapping for a BO or userptr)
42  * All bind operations are implemented via a hybrid approach of using the CPU
56  *	bind BO0 0x0-0x1000
62  *	bind BO1 0x201000-0x202000
66  *	bind BO2 0x1ff000-0x201000
74  * bind can be done immediately (all in-fences satisfied, VM dma-resv kernel
97  * In both modes during the bind IOCTL the user input is validated. In sync
118  * XE_VM_BIND_OP_RESTART operation. When VM async bind work is restarted, the
121  * Bind queues / engines
124  * Think of the case where we have two bind operations A + B and are submitted
125  * in that order. A has in fences while B has none. If using a single bind
128  * limitation by implementing bind engines.
130  * In the bind IOCTL the user can optionally pass in an engine ID which must map
134  * engine's ring. In the example above if A and B have different bind engines B
135  * is free to pass A. If the engine ID field is omitted, the default bind queue
140  * Array of bind operations
144  * of struct drm_xe_vm_bind_op, in a single VM bind IOCTL. This interface
147  * and pass the out fences to the last operation. The ordered nature of a bind
193  * If a VM is in fault mode (TODO: link to fault mode), new bind operations that
297  * When VM is created, a default bind engine and PT table structure are created
326  * the VM bind which will then do the bind immediately.
421  * VM locking protects all of the core data paths (bind operations, execs,
431  * bind path also acquires this lock in write while the exec / compute mode
449  * 1. An exec and bind operation with the same VM can't be executing at the same
452  * 2. A compute mode rebind worker and bind operation with the same VM can't be
492  * of the external BO (if the bind is to an external BO, this is addition to #4)