2014-02-20 17:25:15 +08:00
|
|
|
/*
|
|
|
|
* Copyright 2008 Advanced Micro Devices, Inc.
|
|
|
|
* Copyright 2008 Red Hat Inc.
|
|
|
|
* Copyright 2009 Jerome Glisse.
|
|
|
|
*
|
|
|
|
* Permission is hereby granted, free of charge, to any person obtaining a
|
|
|
|
* copy of this software and associated documentation files (the "Software"),
|
|
|
|
* to deal in the Software without restriction, including without limitation
|
|
|
|
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
|
|
|
|
* and/or sell copies of the Software, and to permit persons to whom the
|
|
|
|
* Software is furnished to do so, subject to the following conditions:
|
|
|
|
*
|
|
|
|
* The above copyright notice and this permission notice shall be included in
|
|
|
|
* all copies or substantial portions of the Software.
|
|
|
|
*
|
|
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
|
|
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
|
|
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
|
|
|
|
* THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
|
|
|
|
* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
|
|
|
|
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
|
|
|
|
* OTHER DEALINGS IN THE SOFTWARE.
|
|
|
|
*
|
|
|
|
* Authors: Dave Airlie
|
|
|
|
* Alex Deucher
|
|
|
|
* Jerome Glisse
|
|
|
|
*/
|
|
|
|
#include <drm/drmP.h>
|
|
|
|
#include <drm/radeon_drm.h>
|
|
|
|
#include "radeon.h"
|
|
|
|
#include "radeon_trace.h"
|
|
|
|
|
|
|
|
/*
|
|
|
|
* GPUVM
|
|
|
|
* GPUVM is similar to the legacy gart on older asics, however
|
|
|
|
* rather than there being a single global gart table
|
|
|
|
* for the entire GPU, there are multiple VM page tables active
|
|
|
|
* at any given time. The VM page tables can contain a mix
|
|
|
|
* vram pages and system memory pages and system memory pages
|
|
|
|
* can be mapped as snooped (cached system pages) or unsnooped
|
|
|
|
* (uncached system pages).
|
|
|
|
* Each VM has an ID associated with it and there is a page table
|
|
|
|
* associated with each VMID. When execting a command buffer,
|
|
|
|
* the kernel tells the the ring what VMID to use for that command
|
|
|
|
* buffer. VMIDs are allocated dynamically as commands are submitted.
|
|
|
|
* The userspace drivers maintain their own address space and the kernel
|
|
|
|
* sets up their pages tables accordingly when they submit their
|
|
|
|
* command buffers and a VMID is assigned.
|
|
|
|
* Cayman/Trinity support up to 8 active VMs at any given time;
|
|
|
|
* SI supports 16.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_num_pde - return the number of page directory entries
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
*
|
|
|
|
* Calculate the number of page directory entries (cayman+).
|
|
|
|
*/
|
|
|
|
static unsigned radeon_vm_num_pdes(struct radeon_device *rdev)
|
|
|
|
{
|
2014-06-06 11:56:50 +08:00
|
|
|
return rdev->vm_manager.max_pfn >> radeon_vm_block_size;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_directory_size - returns the size of the page directory in bytes
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
*
|
|
|
|
* Calculate the size of the page directory in bytes (cayman+).
|
|
|
|
*/
|
|
|
|
static unsigned radeon_vm_directory_size(struct radeon_device *rdev)
|
|
|
|
{
|
|
|
|
return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_manager_init - init the vm manager
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
*
|
|
|
|
* Init the vm manager (cayman+).
|
|
|
|
* Returns 0 for success, error for failure.
|
|
|
|
*/
|
|
|
|
int radeon_vm_manager_init(struct radeon_device *rdev)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
if (!rdev->vm_manager.enabled) {
|
|
|
|
r = radeon_asic_vm_init(rdev);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
rdev->vm_manager.enabled = true;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_manager_fini - tear down the vm manager
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
*
|
|
|
|
* Tear down the VM manager (cayman+).
|
|
|
|
*/
|
|
|
|
void radeon_vm_manager_fini(struct radeon_device *rdev)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!rdev->vm_manager.enabled)
|
|
|
|
return;
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
for (i = 0; i < RADEON_NUM_VM; ++i)
|
2014-02-20 17:25:15 +08:00
|
|
|
radeon_fence_unref(&rdev->vm_manager.active[i]);
|
|
|
|
radeon_asic_vm_fini(rdev);
|
|
|
|
rdev->vm_manager.enabled = false;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2014-02-20 20:42:17 +08:00
|
|
|
* radeon_vm_get_bos - add the vm BOs to a validation list
|
2014-02-20 17:25:15 +08:00
|
|
|
*
|
2014-02-20 20:42:17 +08:00
|
|
|
* @vm: vm providing the BOs
|
|
|
|
* @head: head of validation list
|
2014-02-20 17:25:15 +08:00
|
|
|
*
|
2014-02-20 20:42:17 +08:00
|
|
|
* Add the page directory to the list of BOs to
|
|
|
|
* validate for command submission (cayman+).
|
2014-02-20 17:25:15 +08:00
|
|
|
*/
|
2014-11-27 21:48:42 +08:00
|
|
|
struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
|
2014-03-03 19:38:08 +08:00
|
|
|
struct radeon_vm *vm,
|
|
|
|
struct list_head *head)
|
2014-02-20 17:25:15 +08:00
|
|
|
{
|
2014-11-27 21:48:42 +08:00
|
|
|
struct radeon_bo_list *list;
|
2014-05-28 18:24:17 +08:00
|
|
|
unsigned i, idx;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2017-05-17 20:23:12 +08:00
|
|
|
list = kvmalloc_array(vm->max_pde_used + 2,
|
|
|
|
sizeof(struct radeon_bo_list), GFP_KERNEL);
|
2014-02-20 20:42:17 +08:00
|
|
|
if (!list)
|
|
|
|
return NULL;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
/* add the vm page table to the list */
|
2014-03-03 19:38:08 +08:00
|
|
|
list[0].robj = vm->page_directory;
|
2017-08-08 19:50:46 +08:00
|
|
|
list[0].preferred_domains = RADEON_GEM_DOMAIN_VRAM;
|
2014-06-02 23:33:07 +08:00
|
|
|
list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
|
2014-02-20 20:42:17 +08:00
|
|
|
list[0].tv.bo = &vm->page_directory->tbo;
|
2014-11-19 21:01:23 +08:00
|
|
|
list[0].tv.shared = true;
|
2014-03-03 19:38:08 +08:00
|
|
|
list[0].tiling_flags = 0;
|
2014-02-20 20:42:17 +08:00
|
|
|
list_add(&list[0].tv.head, head);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
for (i = 0, idx = 1; i <= vm->max_pde_used; i++) {
|
|
|
|
if (!vm->page_tables[i].bo)
|
|
|
|
continue;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-03-03 19:38:08 +08:00
|
|
|
list[idx].robj = vm->page_tables[i].bo;
|
2017-08-08 19:50:46 +08:00
|
|
|
list[idx].preferred_domains = RADEON_GEM_DOMAIN_VRAM;
|
2014-06-02 23:33:07 +08:00
|
|
|
list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
|
2014-03-03 19:38:08 +08:00
|
|
|
list[idx].tv.bo = &list[idx].robj->tbo;
|
2014-11-19 21:01:23 +08:00
|
|
|
list[idx].tv.shared = true;
|
2014-03-03 19:38:08 +08:00
|
|
|
list[idx].tiling_flags = 0;
|
2014-02-20 20:42:17 +08:00
|
|
|
list_add(&list[idx++].tv.head, head);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
return list;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_grab_id - allocate the next free VMID
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: vm to allocate id for
|
|
|
|
* @ring: ring we want to submit job to
|
|
|
|
*
|
|
|
|
* Allocate an id for the vm (cayman+).
|
|
|
|
* Returns the fence we need to sync to (if any).
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
|
|
|
struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm, int ring)
|
|
|
|
{
|
|
|
|
struct radeon_fence *best[RADEON_NUM_RINGS] = {};
|
2014-11-19 21:01:25 +08:00
|
|
|
struct radeon_vm_id *vm_id = &vm->ids[ring];
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
unsigned choices[2] = {};
|
|
|
|
unsigned i;
|
|
|
|
|
|
|
|
/* check if the id is still valid */
|
2014-11-19 21:01:25 +08:00
|
|
|
if (vm_id->id && vm_id->last_id_use &&
|
|
|
|
vm_id->last_id_use == rdev->vm_manager.active[vm_id->id])
|
2014-02-20 17:25:15 +08:00
|
|
|
return NULL;
|
|
|
|
|
|
|
|
/* we definately need to flush */
|
2014-11-19 21:01:25 +08:00
|
|
|
vm_id->pd_gpu_addr = ~0ll;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
/* skip over VMID 0, since it is the system VM */
|
|
|
|
for (i = 1; i < rdev->vm_manager.nvm; ++i) {
|
|
|
|
struct radeon_fence *fence = rdev->vm_manager.active[i];
|
|
|
|
|
|
|
|
if (fence == NULL) {
|
|
|
|
/* found a free one */
|
2014-11-19 21:01:25 +08:00
|
|
|
vm_id->id = i;
|
|
|
|
trace_radeon_vm_grab_id(i, ring);
|
2014-02-20 17:25:15 +08:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (radeon_fence_is_earlier(fence, best[fence->ring])) {
|
|
|
|
best[fence->ring] = fence;
|
|
|
|
choices[fence->ring == ring ? 0 : 1] = i;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < 2; ++i) {
|
|
|
|
if (choices[i]) {
|
2014-11-19 21:01:25 +08:00
|
|
|
vm_id->id = choices[i];
|
|
|
|
trace_radeon_vm_grab_id(choices[i], ring);
|
2014-02-20 17:25:15 +08:00
|
|
|
return rdev->vm_manager.active[choices[i]];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* should never happen */
|
|
|
|
BUG();
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2014-02-20 17:47:05 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_flush - hardware flush the vm
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: vm we want to flush
|
|
|
|
* @ring: ring to use for flush
|
2014-11-19 21:01:24 +08:00
|
|
|
* @updates: last vm update that is waited for
|
2014-02-20 17:47:05 +08:00
|
|
|
*
|
|
|
|
* Flush the vm (cayman+).
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
|
|
|
void radeon_vm_flush(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm,
|
2014-11-19 21:01:24 +08:00
|
|
|
int ring, struct radeon_fence *updates)
|
2014-02-20 17:47:05 +08:00
|
|
|
{
|
2014-02-20 20:42:17 +08:00
|
|
|
uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory);
|
2014-11-19 21:01:25 +08:00
|
|
|
struct radeon_vm_id *vm_id = &vm->ids[ring];
|
2014-02-20 20:42:17 +08:00
|
|
|
|
2014-11-19 21:01:25 +08:00
|
|
|
if (pd_addr != vm_id->pd_gpu_addr || !vm_id->flushed_updates ||
|
|
|
|
radeon_fence_is_earlier(vm_id->flushed_updates, updates)) {
|
2014-11-19 21:01:24 +08:00
|
|
|
|
2014-11-19 21:01:25 +08:00
|
|
|
trace_radeon_vm_flush(pd_addr, ring, vm->ids[ring].id);
|
|
|
|
radeon_fence_unref(&vm_id->flushed_updates);
|
|
|
|
vm_id->flushed_updates = radeon_fence_ref(updates);
|
|
|
|
vm_id->pd_gpu_addr = pd_addr;
|
2014-11-19 21:01:19 +08:00
|
|
|
radeon_ring_vm_flush(rdev, &rdev->ring[ring],
|
2014-11-19 21:01:25 +08:00
|
|
|
vm_id->id, vm_id->pd_gpu_addr);
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
}
|
2014-02-20 17:47:05 +08:00
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_fence - remember fence for vm
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: vm we want to fence
|
|
|
|
* @fence: fence to remember
|
|
|
|
*
|
|
|
|
* Fence the vm (cayman+).
|
|
|
|
* Set the fence used to protect page table and id.
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
|
|
|
void radeon_vm_fence(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm,
|
|
|
|
struct radeon_fence *fence)
|
|
|
|
{
|
2014-11-19 21:01:25 +08:00
|
|
|
unsigned vm_id = vm->ids[fence->ring].id;
|
|
|
|
|
|
|
|
radeon_fence_unref(&rdev->vm_manager.active[vm_id]);
|
|
|
|
rdev->vm_manager.active[vm_id] = radeon_fence_ref(fence);
|
2014-02-20 17:47:05 +08:00
|
|
|
|
2014-11-19 21:01:25 +08:00
|
|
|
radeon_fence_unref(&vm->ids[fence->ring].last_id_use);
|
|
|
|
vm->ids[fence->ring].last_id_use = radeon_fence_ref(fence);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_bo_find - find the bo_va for a specific vm & bo
|
|
|
|
*
|
|
|
|
* @vm: requested vm
|
|
|
|
* @bo: requested buffer object
|
|
|
|
*
|
|
|
|
* Find @bo inside the requested vm (cayman+).
|
|
|
|
* Search inside the @bos vm list for the requested vm
|
|
|
|
* Returns the found bo_va or NULL if none is found
|
|
|
|
*
|
|
|
|
* Object has to be reserved!
|
|
|
|
*/
|
|
|
|
struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm,
|
|
|
|
struct radeon_bo *bo)
|
|
|
|
{
|
|
|
|
struct radeon_bo_va *bo_va;
|
|
|
|
|
|
|
|
list_for_each_entry(bo_va, &bo->va, bo_list) {
|
|
|
|
if (bo_va->vm == vm) {
|
|
|
|
return bo_va;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_bo_add - add a bo to a specific vm
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
* @bo: radeon buffer object
|
|
|
|
*
|
|
|
|
* Add @bo into the requested vm (cayman+).
|
|
|
|
* Add @bo to the list of bos associated with the vm
|
|
|
|
* Returns newly added bo_va or NULL for failure
|
|
|
|
*
|
|
|
|
* Object has to be reserved!
|
|
|
|
*/
|
|
|
|
struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm,
|
|
|
|
struct radeon_bo *bo)
|
|
|
|
{
|
|
|
|
struct radeon_bo_va *bo_va;
|
|
|
|
|
|
|
|
bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
|
|
|
|
if (bo_va == NULL) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
bo_va->vm = vm;
|
|
|
|
bo_va->bo = bo;
|
2014-07-30 23:49:56 +08:00
|
|
|
bo_va->it.start = 0;
|
|
|
|
bo_va->it.last = 0;
|
2014-02-20 17:25:15 +08:00
|
|
|
bo_va->flags = 0;
|
|
|
|
bo_va->ref_count = 1;
|
|
|
|
INIT_LIST_HEAD(&bo_va->bo_list);
|
2014-07-18 14:56:40 +08:00
|
|
|
INIT_LIST_HEAD(&bo_va->vm_status);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
mutex_lock(&vm->mutex);
|
|
|
|
list_add_tail(&bo_va->bo_list, &bo->va);
|
|
|
|
mutex_unlock(&vm->mutex);
|
|
|
|
|
|
|
|
return bo_va;
|
|
|
|
}
|
|
|
|
|
2014-07-31 03:05:17 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_set_pages - helper to call the right asic function
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @ib: indirect buffer to fill with commands
|
|
|
|
* @pe: addr of the page entry
|
|
|
|
* @addr: dst addr to write into pe
|
|
|
|
* @count: number of page entries to update
|
|
|
|
* @incr: increase next addr by incr bytes
|
|
|
|
* @flags: hw access flags
|
|
|
|
*
|
|
|
|
* Traces the parameters and calls the right asic functions
|
|
|
|
* to setup the page table using the DMA.
|
|
|
|
*/
|
|
|
|
static void radeon_vm_set_pages(struct radeon_device *rdev,
|
|
|
|
struct radeon_ib *ib,
|
|
|
|
uint64_t pe,
|
|
|
|
uint64_t addr, unsigned count,
|
|
|
|
uint32_t incr, uint32_t flags)
|
|
|
|
{
|
|
|
|
trace_radeon_vm_set_page(pe, addr, count, incr, flags);
|
|
|
|
|
|
|
|
if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
|
|
|
|
uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
|
|
|
|
radeon_asic_vm_copy_pages(rdev, ib, pe, src, count);
|
|
|
|
|
|
|
|
} else if ((flags & R600_PTE_SYSTEM) || (count < 3)) {
|
|
|
|
radeon_asic_vm_write_pages(rdev, ib, pe, addr,
|
|
|
|
count, incr, flags);
|
|
|
|
|
|
|
|
} else {
|
|
|
|
radeon_asic_vm_set_pages(rdev, ib, pe, addr,
|
|
|
|
count, incr, flags);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_clear_bo - initially clear the page dir/table
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @bo: bo to clear
|
|
|
|
*/
|
|
|
|
static int radeon_vm_clear_bo(struct radeon_device *rdev,
|
|
|
|
struct radeon_bo *bo)
|
|
|
|
{
|
|
|
|
struct radeon_ib ib;
|
|
|
|
unsigned entries;
|
|
|
|
uint64_t addr;
|
|
|
|
int r;
|
|
|
|
|
2014-11-19 21:01:23 +08:00
|
|
|
r = radeon_bo_reserve(bo, false);
|
|
|
|
if (r)
|
2014-02-20 20:42:17 +08:00
|
|
|
return r;
|
|
|
|
|
2014-11-19 21:01:23 +08:00
|
|
|
r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
|
|
|
|
if (r)
|
|
|
|
goto error_unreserve;
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
addr = radeon_bo_gpu_offset(bo);
|
|
|
|
entries = radeon_bo_size(bo) / 8;
|
|
|
|
|
2014-07-31 03:05:18 +08:00
|
|
|
r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, 256);
|
2014-02-20 20:42:17 +08:00
|
|
|
if (r)
|
2014-11-19 21:01:23 +08:00
|
|
|
goto error_unreserve;
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
ib.length_dw = 0;
|
|
|
|
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, &ib, addr, 0, entries, 0, 0);
|
|
|
|
radeon_asic_vm_pad_ib(rdev, &ib);
|
2014-07-31 03:05:18 +08:00
|
|
|
WARN_ON(ib.length_dw > 64);
|
2014-02-20 20:42:17 +08:00
|
|
|
|
2014-08-18 16:34:55 +08:00
|
|
|
r = radeon_ib_schedule(rdev, &ib, NULL, false);
|
2014-02-20 20:42:17 +08:00
|
|
|
if (r)
|
2014-11-19 21:01:23 +08:00
|
|
|
goto error_free;
|
2014-02-20 20:42:17 +08:00
|
|
|
|
2014-11-19 21:01:24 +08:00
|
|
|
ib.fence->is_vm_update = true;
|
2014-11-19 21:01:23 +08:00
|
|
|
radeon_bo_fence(bo, ib.fence, false);
|
2014-02-20 20:42:17 +08:00
|
|
|
|
2014-11-19 21:01:23 +08:00
|
|
|
error_free:
|
|
|
|
radeon_ib_free(rdev, &ib);
|
2014-02-20 20:42:17 +08:00
|
|
|
|
2014-11-19 21:01:23 +08:00
|
|
|
error_unreserve:
|
|
|
|
radeon_bo_unreserve(bo);
|
2014-02-20 20:42:17 +08:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_bo_set_addr - set bos virtual address inside a vm
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @bo_va: bo_va to store the address
|
|
|
|
* @soffset: requested offset of the buffer in the VM address space
|
|
|
|
* @flags: attributes of pages (read/write/valid/etc.)
|
|
|
|
*
|
|
|
|
* Set offset of @bo_va (cayman+).
|
|
|
|
* Validate and set the offset requested within the vm address space.
|
|
|
|
* Returns 0 for success, error for failure.
|
|
|
|
*
|
2014-11-19 21:01:20 +08:00
|
|
|
* Object has to be reserved and gets unreserved by this function!
|
2014-02-20 17:25:15 +08:00
|
|
|
*/
|
|
|
|
int radeon_vm_bo_set_addr(struct radeon_device *rdev,
|
|
|
|
struct radeon_bo_va *bo_va,
|
|
|
|
uint64_t soffset,
|
|
|
|
uint32_t flags)
|
|
|
|
{
|
|
|
|
uint64_t size = radeon_bo_size(bo_va->bo);
|
|
|
|
struct radeon_vm *vm = bo_va->vm;
|
2014-02-20 20:42:17 +08:00
|
|
|
unsigned last_pfn, pt_idx;
|
2014-07-30 23:49:56 +08:00
|
|
|
uint64_t eoffset;
|
2014-02-20 20:42:17 +08:00
|
|
|
int r;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
if (soffset) {
|
|
|
|
/* make sure object fit at this offset */
|
2015-11-24 06:39:11 +08:00
|
|
|
eoffset = soffset + size - 1;
|
2014-02-20 17:25:15 +08:00
|
|
|
if (soffset >= eoffset) {
|
2015-06-11 17:38:38 +08:00
|
|
|
r = -EINVAL;
|
|
|
|
goto error_unreserve;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
last_pfn = eoffset / RADEON_GPU_PAGE_SIZE;
|
2015-11-24 06:39:11 +08:00
|
|
|
if (last_pfn >= rdev->vm_manager.max_pfn) {
|
|
|
|
dev_err(rdev->dev, "va above limit (0x%08X >= 0x%08X)\n",
|
2014-02-20 17:25:15 +08:00
|
|
|
last_pfn, rdev->vm_manager.max_pfn);
|
2015-06-11 17:38:38 +08:00
|
|
|
r = -EINVAL;
|
|
|
|
goto error_unreserve;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
} else {
|
|
|
|
eoffset = last_pfn = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
mutex_lock(&vm->mutex);
|
2015-04-27 23:04:36 +08:00
|
|
|
soffset /= RADEON_GPU_PAGE_SIZE;
|
|
|
|
eoffset /= RADEON_GPU_PAGE_SIZE;
|
|
|
|
if (soffset || eoffset) {
|
|
|
|
struct interval_tree_node *it;
|
2015-11-24 06:39:11 +08:00
|
|
|
it = interval_tree_iter_first(&vm->va, soffset, eoffset);
|
2015-04-27 23:04:36 +08:00
|
|
|
if (it && it != &bo_va->it) {
|
|
|
|
struct radeon_bo_va *tmp;
|
|
|
|
tmp = container_of(it, struct radeon_bo_va, it);
|
|
|
|
/* bo and tmp overlap, invalid offset */
|
|
|
|
dev_err(rdev->dev, "bo %p va 0x%010Lx conflict with "
|
|
|
|
"(bo %p 0x%010lx 0x%010lx)\n", bo_va->bo,
|
|
|
|
soffset, tmp->bo, tmp->it.start, tmp->it.last);
|
|
|
|
mutex_unlock(&vm->mutex);
|
2015-06-11 17:38:38 +08:00
|
|
|
r = -EINVAL;
|
|
|
|
goto error_unreserve;
|
2015-04-27 23:04:36 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
if (bo_va->it.start || bo_va->it.last) {
|
2015-07-08 15:56:13 +08:00
|
|
|
/* add a clone of the bo_va to clear the old address */
|
|
|
|
struct radeon_bo_va *tmp;
|
|
|
|
tmp = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
|
|
|
|
if (!tmp) {
|
|
|
|
mutex_unlock(&vm->mutex);
|
|
|
|
r = -ENOMEM;
|
|
|
|
goto error_unreserve;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
2015-07-08 15:56:13 +08:00
|
|
|
tmp->it.start = bo_va->it.start;
|
|
|
|
tmp->it.last = bo_va->it.last;
|
|
|
|
tmp->vm = vm;
|
|
|
|
tmp->bo = radeon_bo_ref(bo_va->bo);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
interval_tree_remove(&bo_va->it, &vm->va);
|
2015-07-08 15:56:14 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
2014-07-30 23:49:56 +08:00
|
|
|
bo_va->it.start = 0;
|
|
|
|
bo_va->it.last = 0;
|
2015-07-08 15:56:13 +08:00
|
|
|
list_del_init(&bo_va->vm_status);
|
|
|
|
list_add(&tmp->vm_status, &vm->freed);
|
|
|
|
spin_unlock(&vm->status_lock);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
if (soffset || eoffset) {
|
2015-07-08 15:56:14 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
2014-07-30 23:49:56 +08:00
|
|
|
bo_va->it.start = soffset;
|
2015-11-24 06:39:11 +08:00
|
|
|
bo_va->it.last = eoffset;
|
2015-05-26 18:24:15 +08:00
|
|
|
list_add(&bo_va->vm_status, &vm->cleared);
|
|
|
|
spin_unlock(&vm->status_lock);
|
2015-07-08 15:56:14 +08:00
|
|
|
interval_tree_insert(&bo_va->it, &vm->va);
|
2014-07-18 14:56:40 +08:00
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
bo_va->flags = flags;
|
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
soffset >>= radeon_vm_block_size;
|
|
|
|
eoffset >>= radeon_vm_block_size;
|
2014-06-06 11:56:50 +08:00
|
|
|
|
|
|
|
BUG_ON(eoffset >= radeon_vm_num_pdes(rdev));
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
if (eoffset > vm->max_pde_used)
|
|
|
|
vm->max_pde_used = eoffset;
|
|
|
|
|
|
|
|
radeon_bo_unreserve(bo_va->bo);
|
|
|
|
|
|
|
|
/* walk over the address space and allocate the page tables */
|
|
|
|
for (pt_idx = soffset; pt_idx <= eoffset; ++pt_idx) {
|
|
|
|
struct radeon_bo *pt;
|
|
|
|
|
|
|
|
if (vm->page_tables[pt_idx].bo)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
/* drop mutex to allocate and clear page table */
|
|
|
|
mutex_unlock(&vm->mutex);
|
|
|
|
|
|
|
|
r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8,
|
2014-07-03 03:28:10 +08:00
|
|
|
RADEON_GPU_PAGE_SIZE, true,
|
2014-09-18 20:11:56 +08:00
|
|
|
RADEON_GEM_DOMAIN_VRAM, 0,
|
|
|
|
NULL, NULL, &pt);
|
2014-02-20 20:42:17 +08:00
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
r = radeon_vm_clear_bo(rdev, pt);
|
|
|
|
if (r) {
|
|
|
|
radeon_bo_unref(&pt);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* aquire mutex again */
|
|
|
|
mutex_lock(&vm->mutex);
|
|
|
|
if (vm->page_tables[pt_idx].bo) {
|
|
|
|
/* someone else allocated the pt in the meantime */
|
|
|
|
mutex_unlock(&vm->mutex);
|
|
|
|
radeon_bo_unref(&pt);
|
|
|
|
mutex_lock(&vm->mutex);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
vm->page_tables[pt_idx].addr = 0;
|
|
|
|
vm->page_tables[pt_idx].bo = pt;
|
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
mutex_unlock(&vm->mutex);
|
2014-11-19 21:01:20 +08:00
|
|
|
return 0;
|
2015-06-11 17:38:38 +08:00
|
|
|
|
|
|
|
error_unreserve:
|
|
|
|
radeon_bo_unreserve(bo_va->bo);
|
|
|
|
return r;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_map_gart - get the physical address of a gart page
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @addr: the unmapped addr
|
|
|
|
*
|
|
|
|
* Look up the physical address of the page that the pte resolves
|
|
|
|
* to (cayman+).
|
|
|
|
* Returns the physical address of the page.
|
|
|
|
*/
|
|
|
|
uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr)
|
|
|
|
{
|
|
|
|
uint64_t result;
|
|
|
|
|
|
|
|
/* page table offset */
|
2015-01-21 16:36:37 +08:00
|
|
|
result = rdev->gart.pages_entry[addr >> RADEON_GPU_PAGE_SHIFT];
|
|
|
|
result &= ~RADEON_GPU_PAGE_MASK;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_page_flags - translate page flags to what the hw uses
|
|
|
|
*
|
|
|
|
* @flags: flags comming from userspace
|
|
|
|
*
|
|
|
|
* Translate the flags the userspace ABI uses to hw flags.
|
|
|
|
*/
|
|
|
|
static uint32_t radeon_vm_page_flags(uint32_t flags)
|
|
|
|
{
|
2016-03-16 19:56:45 +08:00
|
|
|
uint32_t hw_flags = 0;
|
|
|
|
|
|
|
|
hw_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0;
|
|
|
|
hw_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0;
|
|
|
|
hw_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0;
|
|
|
|
if (flags & RADEON_VM_PAGE_SYSTEM) {
|
|
|
|
hw_flags |= R600_PTE_SYSTEM;
|
|
|
|
hw_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0;
|
|
|
|
}
|
|
|
|
return hw_flags;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_update_pdes - make sure that page directory is valid
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
* @start: start of GPU address range
|
|
|
|
* @end: end of GPU address range
|
|
|
|
*
|
|
|
|
* Allocates new page tables if necessary
|
|
|
|
* and updates the page directory (cayman+).
|
|
|
|
* Returns 0 for success, error for failure.
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
2014-02-20 20:42:17 +08:00
|
|
|
int radeon_vm_update_page_directory(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm)
|
2014-02-20 17:25:15 +08:00
|
|
|
{
|
2014-05-30 21:21:16 +08:00
|
|
|
struct radeon_bo *pd = vm->page_directory;
|
|
|
|
uint64_t pd_addr = radeon_bo_gpu_offset(pd);
|
2014-06-06 11:56:50 +08:00
|
|
|
uint32_t incr = RADEON_VM_PTE_COUNT * 8;
|
2014-02-20 17:25:15 +08:00
|
|
|
uint64_t last_pde = ~0, last_pt = ~0;
|
2014-02-20 20:42:17 +08:00
|
|
|
unsigned count = 0, pt_idx, ndw;
|
|
|
|
struct radeon_ib ib;
|
2014-02-20 17:25:15 +08:00
|
|
|
int r;
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
/* padding, etc. */
|
|
|
|
ndw = 64;
|
|
|
|
|
|
|
|
/* assume the worst case */
|
2014-07-31 03:05:18 +08:00
|
|
|
ndw += vm->max_pde_used * 6;
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
/* update too big for an IB */
|
|
|
|
if (ndw > 0xfffff)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
ib.length_dw = 0;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
/* walk over the address space and update the page directory */
|
2014-02-20 20:42:17 +08:00
|
|
|
for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) {
|
|
|
|
struct radeon_bo *bo = vm->page_tables[pt_idx].bo;
|
2014-02-20 17:25:15 +08:00
|
|
|
uint64_t pde, pt;
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
if (bo == NULL)
|
2014-02-20 17:25:15 +08:00
|
|
|
continue;
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
pt = radeon_bo_gpu_offset(bo);
|
|
|
|
if (vm->page_tables[pt_idx].addr == pt)
|
|
|
|
continue;
|
|
|
|
vm->page_tables[pt_idx].addr = pt;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
pde = pd_addr + pt_idx * 8;
|
2014-02-20 17:25:15 +08:00
|
|
|
if (((last_pde + 8 * count) != pde) ||
|
|
|
|
((last_pt + incr * count) != pt)) {
|
|
|
|
|
|
|
|
if (count) {
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, &ib, last_pde,
|
|
|
|
last_pt, count, incr,
|
|
|
|
R600_PTE_VALID);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
count = 1;
|
|
|
|
last_pde = pde;
|
|
|
|
last_pt = pt;
|
|
|
|
} else {
|
|
|
|
++count;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
if (count)
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, &ib, last_pde, last_pt, count,
|
|
|
|
incr, R600_PTE_VALID);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-02-20 20:42:17 +08:00
|
|
|
if (ib.length_dw != 0) {
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_asic_vm_pad_ib(rdev, &ib);
|
2014-04-02 23:14:48 +08:00
|
|
|
|
2014-11-19 21:01:27 +08:00
|
|
|
radeon_sync_resv(rdev, &ib.sync, pd->tbo.resv, true);
|
2014-07-31 03:05:18 +08:00
|
|
|
WARN_ON(ib.length_dw > ndw);
|
2014-08-18 16:34:55 +08:00
|
|
|
r = radeon_ib_schedule(rdev, &ib, NULL, false);
|
2014-02-20 20:42:17 +08:00
|
|
|
if (r) {
|
|
|
|
radeon_ib_free(rdev, &ib);
|
|
|
|
return r;
|
|
|
|
}
|
2014-11-19 21:01:24 +08:00
|
|
|
ib.fence->is_vm_update = true;
|
2014-11-19 21:01:23 +08:00
|
|
|
radeon_bo_fence(pd, ib.fence, false);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
2014-02-20 20:42:17 +08:00
|
|
|
radeon_ib_free(rdev, &ib);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-05-10 18:17:55 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_frag_ptes - add fragment information to PTEs
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @ib: IB for the update
|
|
|
|
* @pe_start: first PTE to handle
|
|
|
|
* @pe_end: last PTE to handle
|
|
|
|
* @addr: addr those PTEs should point to
|
|
|
|
* @flags: hw mapping flags
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
|
|
|
static void radeon_vm_frag_ptes(struct radeon_device *rdev,
|
|
|
|
struct radeon_ib *ib,
|
|
|
|
uint64_t pe_start, uint64_t pe_end,
|
|
|
|
uint64_t addr, uint32_t flags)
|
|
|
|
{
|
|
|
|
/**
|
|
|
|
* The MC L1 TLB supports variable sized pages, based on a fragment
|
|
|
|
* field in the PTE. When this field is set to a non-zero value, page
|
|
|
|
* granularity is increased from 4KB to (1 << (12 + frag)). The PTE
|
|
|
|
* flags are considered valid for all PTEs within the fragment range
|
|
|
|
* and corresponding mappings are assumed to be physically contiguous.
|
|
|
|
*
|
|
|
|
* The L1 TLB can store a single PTE for the whole fragment,
|
|
|
|
* significantly increasing the space available for translation
|
|
|
|
* caching. This leads to large improvements in throughput when the
|
|
|
|
* TLB is under pressure.
|
|
|
|
*
|
|
|
|
* The L2 TLB distributes small and large fragments into two
|
|
|
|
* asymmetric partitions. The large fragment cache is significantly
|
|
|
|
* larger. Thus, we try to use large fragments wherever possible.
|
|
|
|
* Userspace can support this by aligning virtual base address and
|
|
|
|
* allocation size to the fragment size.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* NI is optimized for 256KB fragments, SI and newer for 64KB */
|
2015-01-30 22:52:24 +08:00
|
|
|
uint64_t frag_flags = ((rdev->family == CHIP_CAYMAN) ||
|
|
|
|
(rdev->family == CHIP_ARUBA)) ?
|
2014-05-10 18:17:55 +08:00
|
|
|
R600_PTE_FRAG_256KB : R600_PTE_FRAG_64KB;
|
2015-01-30 22:52:24 +08:00
|
|
|
uint64_t frag_align = ((rdev->family == CHIP_CAYMAN) ||
|
|
|
|
(rdev->family == CHIP_ARUBA)) ? 0x200 : 0x80;
|
2014-05-10 18:17:55 +08:00
|
|
|
|
|
|
|
uint64_t frag_start = ALIGN(pe_start, frag_align);
|
|
|
|
uint64_t frag_end = pe_end & ~(frag_align - 1);
|
|
|
|
|
|
|
|
unsigned count;
|
|
|
|
|
|
|
|
/* system pages are non continuously */
|
|
|
|
if ((flags & R600_PTE_SYSTEM) || !(flags & R600_PTE_VALID) ||
|
|
|
|
(frag_start >= frag_end)) {
|
|
|
|
|
|
|
|
count = (pe_end - pe_start) / 8;
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, ib, pe_start, addr, count,
|
|
|
|
RADEON_GPU_PAGE_SIZE, flags);
|
2014-05-10 18:17:55 +08:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* handle the 4K area at the beginning */
|
|
|
|
if (pe_start != frag_start) {
|
|
|
|
count = (frag_start - pe_start) / 8;
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, ib, pe_start, addr, count,
|
|
|
|
RADEON_GPU_PAGE_SIZE, flags);
|
2014-05-10 18:17:55 +08:00
|
|
|
addr += RADEON_GPU_PAGE_SIZE * count;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* handle the area in the middle */
|
|
|
|
count = (frag_end - frag_start) / 8;
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, ib, frag_start, addr, count,
|
|
|
|
RADEON_GPU_PAGE_SIZE, flags | frag_flags);
|
2014-05-10 18:17:55 +08:00
|
|
|
|
|
|
|
/* handle the 4K area at the end */
|
|
|
|
if (frag_end != pe_end) {
|
|
|
|
addr += RADEON_GPU_PAGE_SIZE * count;
|
|
|
|
count = (pe_end - frag_end) / 8;
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_vm_set_pages(rdev, ib, frag_end, addr, count,
|
|
|
|
RADEON_GPU_PAGE_SIZE, flags);
|
2014-05-10 18:17:55 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_update_ptes - make sure that page tables are valid
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
* @start: start of GPU address range
|
|
|
|
* @end: end of GPU address range
|
|
|
|
* @dst: destination address to map to
|
|
|
|
* @flags: mapping flags
|
|
|
|
*
|
|
|
|
* Update the page tables in the range @start - @end (cayman+).
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
2014-11-27 21:48:43 +08:00
|
|
|
static int radeon_vm_update_ptes(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm,
|
|
|
|
struct radeon_ib *ib,
|
|
|
|
uint64_t start, uint64_t end,
|
|
|
|
uint64_t dst, uint32_t flags)
|
2014-02-20 17:25:15 +08:00
|
|
|
{
|
2014-06-06 11:56:50 +08:00
|
|
|
uint64_t mask = RADEON_VM_PTE_COUNT - 1;
|
2014-02-20 17:25:15 +08:00
|
|
|
uint64_t last_pte = ~0, last_dst = ~0;
|
|
|
|
unsigned count = 0;
|
|
|
|
uint64_t addr;
|
|
|
|
|
|
|
|
/* walk over the address space and update the page tables */
|
|
|
|
for (addr = start; addr < end; ) {
|
2014-06-06 11:56:50 +08:00
|
|
|
uint64_t pt_idx = addr >> radeon_vm_block_size;
|
2014-05-30 21:21:16 +08:00
|
|
|
struct radeon_bo *pt = vm->page_tables[pt_idx].bo;
|
2014-02-20 17:25:15 +08:00
|
|
|
unsigned nptes;
|
|
|
|
uint64_t pte;
|
2014-11-27 21:48:43 +08:00
|
|
|
int r;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-11-19 21:01:28 +08:00
|
|
|
radeon_sync_resv(rdev, &ib->sync, pt->tbo.resv, true);
|
2014-11-27 21:48:43 +08:00
|
|
|
r = reservation_object_reserve_shared(pt->tbo.resv);
|
|
|
|
if (r)
|
|
|
|
return r;
|
2014-05-30 21:21:16 +08:00
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
if ((addr & ~mask) == (end & ~mask))
|
|
|
|
nptes = end - addr;
|
|
|
|
else
|
|
|
|
nptes = RADEON_VM_PTE_COUNT - (addr & mask);
|
|
|
|
|
2014-05-30 21:21:16 +08:00
|
|
|
pte = radeon_bo_gpu_offset(pt);
|
2014-02-20 17:25:15 +08:00
|
|
|
pte += (addr & mask) * 8;
|
|
|
|
|
|
|
|
if ((last_pte + 8 * count) != pte) {
|
|
|
|
|
|
|
|
if (count) {
|
2014-05-10 18:17:55 +08:00
|
|
|
radeon_vm_frag_ptes(rdev, ib, last_pte,
|
|
|
|
last_pte + 8 * count,
|
|
|
|
last_dst, flags);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
count = nptes;
|
|
|
|
last_pte = pte;
|
|
|
|
last_dst = dst;
|
|
|
|
} else {
|
|
|
|
count += nptes;
|
|
|
|
}
|
|
|
|
|
|
|
|
addr += nptes;
|
|
|
|
dst += nptes * RADEON_GPU_PAGE_SIZE;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (count) {
|
2014-05-10 18:17:55 +08:00
|
|
|
radeon_vm_frag_ptes(rdev, ib, last_pte,
|
|
|
|
last_pte + 8 * count,
|
|
|
|
last_dst, flags);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
2014-11-27 21:48:43 +08:00
|
|
|
|
|
|
|
return 0;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
2014-11-19 21:01:23 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_fence_pts - fence page tables after an update
|
|
|
|
*
|
|
|
|
* @vm: requested vm
|
|
|
|
* @start: start of GPU address range
|
|
|
|
* @end: end of GPU address range
|
|
|
|
* @fence: fence to use
|
|
|
|
*
|
|
|
|
* Fence the page tables in the range @start - @end (cayman+).
|
|
|
|
*
|
|
|
|
* Global and local mutex must be locked!
|
|
|
|
*/
|
|
|
|
static void radeon_vm_fence_pts(struct radeon_vm *vm,
|
|
|
|
uint64_t start, uint64_t end,
|
|
|
|
struct radeon_fence *fence)
|
|
|
|
{
|
|
|
|
unsigned i;
|
|
|
|
|
|
|
|
start >>= radeon_vm_block_size;
|
2015-11-24 06:39:11 +08:00
|
|
|
end = (end - 1) >> radeon_vm_block_size;
|
2014-11-19 21:01:23 +08:00
|
|
|
|
|
|
|
for (i = start; i <= end; ++i)
|
2014-11-27 21:48:43 +08:00
|
|
|
radeon_bo_fence(vm->page_tables[i].bo, fence, true);
|
2014-11-19 21:01:23 +08:00
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_bo_update - map a bo into the vm page table
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
* @bo: radeon buffer object
|
|
|
|
* @mem: ttm mem
|
|
|
|
*
|
|
|
|
* Fill in the page table entries for @bo (cayman+).
|
|
|
|
* Returns 0 for success, -EINVAL for failure.
|
|
|
|
*
|
2014-02-21 02:33:15 +08:00
|
|
|
* Object have to be reserved and mutex must be locked!
|
2014-02-20 17:25:15 +08:00
|
|
|
*/
|
|
|
|
int radeon_vm_bo_update(struct radeon_device *rdev,
|
2014-07-18 14:56:40 +08:00
|
|
|
struct radeon_bo_va *bo_va,
|
2014-02-20 17:25:15 +08:00
|
|
|
struct ttm_mem_reg *mem)
|
|
|
|
{
|
2014-07-18 14:56:40 +08:00
|
|
|
struct radeon_vm *vm = bo_va->vm;
|
2014-02-20 17:25:15 +08:00
|
|
|
struct radeon_ib ib;
|
2014-07-31 03:05:18 +08:00
|
|
|
unsigned nptes, ncmds, ndw;
|
2014-02-20 17:25:15 +08:00
|
|
|
uint64_t addr;
|
2014-07-31 03:05:18 +08:00
|
|
|
uint32_t flags;
|
2014-02-20 17:25:15 +08:00
|
|
|
int r;
|
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
if (!bo_va->it.start) {
|
2014-02-20 17:25:15 +08:00
|
|
|
dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n",
|
2014-07-18 14:56:40 +08:00
|
|
|
bo_va->bo, vm);
|
2014-02-20 17:25:15 +08:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
2015-05-26 18:24:15 +08:00
|
|
|
if (mem) {
|
|
|
|
if (list_empty(&bo_va->vm_status)) {
|
|
|
|
spin_unlock(&vm->status_lock);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
list_del_init(&bo_va->vm_status);
|
|
|
|
} else {
|
|
|
|
list_del(&bo_va->vm_status);
|
|
|
|
list_add(&bo_va->vm_status, &vm->cleared);
|
|
|
|
}
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_unlock(&vm->status_lock);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
bo_va->flags &= ~RADEON_VM_PAGE_VALID;
|
|
|
|
bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
|
2014-07-17 18:01:08 +08:00
|
|
|
bo_va->flags &= ~RADEON_VM_PAGE_SNOOPED;
|
drm/radeon: add userptr support v8
This patch adds an IOCTL for turning a pointer supplied by
userspace into a buffer object.
It imposes several restrictions upon the memory being mapped:
1. It must be page aligned (both start/end addresses, i.e ptr and size).
2. It must be normal system memory, not a pointer into another map of IO
space (e.g. it must not be a GTT mmapping of another object).
3. The BO is mapped into GTT, so the maximum amount of memory mapped at
all times is still the GTT limit.
4. The BO is only mapped readonly for now, so no write support.
5. List of backing pages is only acquired once, so they represent a
snapshot of the first use.
Exporting and sharing as well as mapping of buffer objects created by
this function is forbidden and results in an -EPERM.
v2: squash all previous changes into first public version
v3: fix tabs, map readonly, don't use MM callback any more
v4: set TTM_PAGE_FLAG_SG so that TTM never messes with the pages,
pin/unpin pages on bind/unbind instead of populate/unpopulate
v5: rebased on 3.17-wip, IOCTL renamed to userptr, reject any unknown
flags, better handle READONLY flag, improve permission check
v6: fix ptr cast warning, use set_page_dirty/mark_page_accessed on unpin
v7: add warning about it's availability in the API definition
v8: drop access_ok check, fix VM mapping bits
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v4)
Reviewed-by: Jérôme Glisse <jglisse@redhat.com> (v4)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2014-08-07 15:36:00 +08:00
|
|
|
if (bo_va->bo && radeon_ttm_tt_is_readonly(bo_va->bo->tbo.ttm))
|
|
|
|
bo_va->flags &= ~RADEON_VM_PAGE_WRITEABLE;
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
if (mem) {
|
|
|
|
addr = mem->start << PAGE_SHIFT;
|
|
|
|
if (mem->mem_type != TTM_PL_SYSTEM) {
|
|
|
|
bo_va->flags |= RADEON_VM_PAGE_VALID;
|
|
|
|
}
|
|
|
|
if (mem->mem_type == TTM_PL_TT) {
|
|
|
|
bo_va->flags |= RADEON_VM_PAGE_SYSTEM;
|
2014-07-17 18:01:08 +08:00
|
|
|
if (!(bo_va->bo->flags & (RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC)))
|
|
|
|
bo_va->flags |= RADEON_VM_PAGE_SNOOPED;
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
} else {
|
|
|
|
addr += rdev->vm_manager.vram_base_offset;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
addr = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
trace_radeon_vm_bo_update(bo_va);
|
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
nptes = bo_va->it.last - bo_va->it.start + 1;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-07-31 03:05:18 +08:00
|
|
|
/* reserve space for one command every (1 << BLOCK_SIZE) entries
|
|
|
|
or 2k dwords (whatever is smaller) */
|
|
|
|
ncmds = (nptes >> min(radeon_vm_block_size, 11)) + 1;
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
/* padding, etc. */
|
|
|
|
ndw = 64;
|
|
|
|
|
2014-07-31 03:05:18 +08:00
|
|
|
flags = radeon_vm_page_flags(bo_va->flags);
|
|
|
|
if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
|
|
|
|
/* only copy commands needed */
|
|
|
|
ndw += ncmds * 7;
|
|
|
|
|
|
|
|
} else if (flags & R600_PTE_SYSTEM) {
|
|
|
|
/* header for write data commands */
|
|
|
|
ndw += ncmds * 4;
|
|
|
|
|
|
|
|
/* body of write data command */
|
|
|
|
ndw += nptes * 2;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-07-31 03:05:18 +08:00
|
|
|
} else {
|
|
|
|
/* set page commands needed */
|
|
|
|
ndw += ncmds * 10;
|
|
|
|
|
|
|
|
/* two extra commands for begin/end of fragment */
|
|
|
|
ndw += 2 * 10;
|
|
|
|
}
|
2014-02-20 17:25:15 +08:00
|
|
|
|
|
|
|
/* update too big for an IB */
|
|
|
|
if (ndw > 0xfffff)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
ib.length_dw = 0;
|
|
|
|
|
2014-11-19 21:01:28 +08:00
|
|
|
if (!(bo_va->flags & RADEON_VM_PAGE_VALID)) {
|
|
|
|
unsigned i;
|
|
|
|
|
|
|
|
for (i = 0; i < RADEON_NUM_RINGS; ++i)
|
|
|
|
radeon_sync_fence(&ib.sync, vm->ids[i].last_id_use);
|
|
|
|
}
|
|
|
|
|
2014-11-27 21:48:43 +08:00
|
|
|
r = radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
|
|
|
|
bo_va->it.last + 1, addr,
|
|
|
|
radeon_vm_page_flags(bo_va->flags));
|
|
|
|
if (r) {
|
|
|
|
radeon_ib_free(rdev, &ib);
|
|
|
|
return r;
|
|
|
|
}
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-07-31 03:05:17 +08:00
|
|
|
radeon_asic_vm_pad_ib(rdev, &ib);
|
2014-07-31 03:05:18 +08:00
|
|
|
WARN_ON(ib.length_dw > ndw);
|
|
|
|
|
2014-08-18 16:34:55 +08:00
|
|
|
r = radeon_ib_schedule(rdev, &ib, NULL, false);
|
2014-02-20 17:25:15 +08:00
|
|
|
if (r) {
|
|
|
|
radeon_ib_free(rdev, &ib);
|
|
|
|
return r;
|
|
|
|
}
|
2014-11-19 21:01:24 +08:00
|
|
|
ib.fence->is_vm_update = true;
|
2014-11-19 21:01:23 +08:00
|
|
|
radeon_vm_fence_pts(vm, bo_va->it.start, bo_va->it.last + 1, ib.fence);
|
2014-11-19 21:01:26 +08:00
|
|
|
radeon_fence_unref(&bo_va->last_pt_update);
|
|
|
|
bo_va->last_pt_update = radeon_fence_ref(ib.fence);
|
2014-02-20 17:25:15 +08:00
|
|
|
radeon_ib_free(rdev, &ib);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-07-18 14:56:40 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_clear_freed - clear freed BOs in the PT
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
*
|
|
|
|
* Make sure all freed BOs are cleared in the PT.
|
|
|
|
* Returns 0 for success.
|
|
|
|
*
|
|
|
|
* PTs have to be reserved and mutex must be locked!
|
|
|
|
*/
|
|
|
|
int radeon_vm_clear_freed(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm)
|
|
|
|
{
|
2014-11-27 21:48:44 +08:00
|
|
|
struct radeon_bo_va *bo_va;
|
2015-05-26 18:24:15 +08:00
|
|
|
int r = 0;
|
2014-07-18 14:56:40 +08:00
|
|
|
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
|
|
|
while (!list_empty(&vm->freed)) {
|
|
|
|
bo_va = list_first_entry(&vm->freed,
|
|
|
|
struct radeon_bo_va, vm_status);
|
|
|
|
spin_unlock(&vm->status_lock);
|
|
|
|
|
2014-07-18 14:56:40 +08:00
|
|
|
r = radeon_vm_bo_update(rdev, bo_va, NULL);
|
2014-07-31 03:04:57 +08:00
|
|
|
radeon_bo_unref(&bo_va->bo);
|
2014-11-19 21:01:26 +08:00
|
|
|
radeon_fence_unref(&bo_va->last_pt_update);
|
2015-05-26 18:24:15 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
|
|
|
list_del(&bo_va->vm_status);
|
2014-07-18 14:56:40 +08:00
|
|
|
kfree(bo_va);
|
|
|
|
if (r)
|
2015-05-26 18:24:15 +08:00
|
|
|
break;
|
2014-11-27 21:48:44 +08:00
|
|
|
|
2014-07-18 14:56:40 +08:00
|
|
|
}
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_unlock(&vm->status_lock);
|
2015-05-26 18:24:15 +08:00
|
|
|
return r;
|
2014-07-18 14:56:40 +08:00
|
|
|
|
|
|
|
}
|
|
|
|
|
2014-07-18 15:24:53 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_clear_invalids - clear invalidated BOs in the PT
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
*
|
|
|
|
* Make sure all invalidated BOs are cleared in the PT.
|
|
|
|
* Returns 0 for success.
|
|
|
|
*
|
|
|
|
* PTs have to be reserved and mutex must be locked!
|
|
|
|
*/
|
|
|
|
int radeon_vm_clear_invalids(struct radeon_device *rdev,
|
|
|
|
struct radeon_vm *vm)
|
|
|
|
{
|
2014-11-27 21:48:44 +08:00
|
|
|
struct radeon_bo_va *bo_va;
|
2014-07-18 15:24:53 +08:00
|
|
|
int r;
|
|
|
|
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
|
|
|
while (!list_empty(&vm->invalidated)) {
|
|
|
|
bo_va = list_first_entry(&vm->invalidated,
|
|
|
|
struct radeon_bo_va, vm_status);
|
|
|
|
spin_unlock(&vm->status_lock);
|
|
|
|
|
2014-07-18 15:24:53 +08:00
|
|
|
r = radeon_vm_bo_update(rdev, bo_va, NULL);
|
|
|
|
if (r)
|
|
|
|
return r;
|
2014-11-27 21:48:44 +08:00
|
|
|
|
|
|
|
spin_lock(&vm->status_lock);
|
2014-07-18 15:24:53 +08:00
|
|
|
}
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_unlock(&vm->status_lock);
|
|
|
|
|
2014-07-18 15:24:53 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-02-20 17:25:15 +08:00
|
|
|
/**
|
|
|
|
* radeon_vm_bo_rmv - remove a bo to a specific vm
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @bo_va: requested bo_va
|
|
|
|
*
|
|
|
|
* Remove @bo_va->bo from the requested vm (cayman+).
|
|
|
|
*
|
|
|
|
* Object have to be reserved!
|
|
|
|
*/
|
2014-07-18 14:56:40 +08:00
|
|
|
void radeon_vm_bo_rmv(struct radeon_device *rdev,
|
|
|
|
struct radeon_bo_va *bo_va)
|
2014-02-20 17:25:15 +08:00
|
|
|
{
|
2014-07-18 14:56:40 +08:00
|
|
|
struct radeon_vm *vm = bo_va->vm;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-07-18 14:56:40 +08:00
|
|
|
list_del(&bo_va->bo_list);
|
2014-02-21 02:33:15 +08:00
|
|
|
|
2014-07-18 14:56:40 +08:00
|
|
|
mutex_lock(&vm->mutex);
|
2015-04-27 23:04:34 +08:00
|
|
|
if (bo_va->it.start || bo_va->it.last)
|
|
|
|
interval_tree_remove(&bo_va->it, &vm->va);
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2015-05-26 18:24:15 +08:00
|
|
|
spin_lock(&vm->status_lock);
|
2015-06-26 18:31:29 +08:00
|
|
|
list_del(&bo_va->vm_status);
|
|
|
|
if (bo_va->it.start || bo_va->it.last) {
|
2014-07-31 03:04:57 +08:00
|
|
|
bo_va->bo = radeon_bo_ref(bo_va->bo);
|
2014-07-18 14:56:40 +08:00
|
|
|
list_add(&bo_va->vm_status, &vm->freed);
|
|
|
|
} else {
|
2014-11-19 21:01:26 +08:00
|
|
|
radeon_fence_unref(&bo_va->last_pt_update);
|
2014-07-18 14:56:40 +08:00
|
|
|
kfree(bo_va);
|
|
|
|
}
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_unlock(&vm->status_lock);
|
2014-07-18 14:56:40 +08:00
|
|
|
|
|
|
|
mutex_unlock(&vm->mutex);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_bo_invalidate - mark the bo as invalid
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
* @bo: radeon buffer object
|
|
|
|
*
|
|
|
|
* Mark @bo as invalid (cayman+).
|
|
|
|
*/
|
|
|
|
void radeon_vm_bo_invalidate(struct radeon_device *rdev,
|
|
|
|
struct radeon_bo *bo)
|
|
|
|
{
|
|
|
|
struct radeon_bo_va *bo_va;
|
|
|
|
|
|
|
|
list_for_each_entry(bo_va, &bo->va, bo_list) {
|
2015-05-26 18:24:15 +08:00
|
|
|
spin_lock(&bo_va->vm->status_lock);
|
2015-07-08 15:56:14 +08:00
|
|
|
if (list_empty(&bo_va->vm_status) &&
|
|
|
|
(bo_va->it.start || bo_va->it.last))
|
2014-07-18 15:24:53 +08:00
|
|
|
list_add(&bo_va->vm_status, &bo_va->vm->invalidated);
|
2015-05-26 18:24:15 +08:00
|
|
|
spin_unlock(&bo_va->vm->status_lock);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_init - initialize a vm instance
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
*
|
|
|
|
* Init @vm fields (cayman+).
|
|
|
|
*/
|
2014-02-20 20:42:17 +08:00
|
|
|
int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
|
2014-02-20 17:25:15 +08:00
|
|
|
{
|
2014-05-10 18:17:56 +08:00
|
|
|
const unsigned align = min(RADEON_VM_PTB_ALIGN_SIZE,
|
|
|
|
RADEON_VM_PTE_COUNT * 8);
|
2014-02-20 20:42:17 +08:00
|
|
|
unsigned pd_size, pd_entries, pts_size;
|
2014-11-19 21:01:25 +08:00
|
|
|
int i, r;
|
2014-02-20 20:42:17 +08:00
|
|
|
|
2014-07-18 19:48:10 +08:00
|
|
|
vm->ib_bo_va = NULL;
|
2014-11-19 21:01:25 +08:00
|
|
|
for (i = 0; i < RADEON_NUM_RINGS; ++i) {
|
|
|
|
vm->ids[i].id = 0;
|
|
|
|
vm->ids[i].flushed_updates = NULL;
|
|
|
|
vm->ids[i].last_id_use = NULL;
|
|
|
|
}
|
2014-02-20 17:25:15 +08:00
|
|
|
mutex_init(&vm->mutex);
|
2014-07-30 23:49:56 +08:00
|
|
|
vm->va = RB_ROOT;
|
2014-11-27 21:48:44 +08:00
|
|
|
spin_lock_init(&vm->status_lock);
|
2014-07-18 15:24:53 +08:00
|
|
|
INIT_LIST_HEAD(&vm->invalidated);
|
2014-07-18 14:56:40 +08:00
|
|
|
INIT_LIST_HEAD(&vm->freed);
|
2015-05-26 18:24:15 +08:00
|
|
|
INIT_LIST_HEAD(&vm->cleared);
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
pd_size = radeon_vm_directory_size(rdev);
|
|
|
|
pd_entries = radeon_vm_num_pdes(rdev);
|
|
|
|
|
|
|
|
/* allocate page table array */
|
|
|
|
pts_size = pd_entries * sizeof(struct radeon_vm_pt);
|
|
|
|
vm->page_tables = kzalloc(pts_size, GFP_KERNEL);
|
|
|
|
if (vm->page_tables == NULL) {
|
|
|
|
DRM_ERROR("Cannot allocate memory for page table array\n");
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2014-07-03 03:28:10 +08:00
|
|
|
r = radeon_bo_create(rdev, pd_size, align, true,
|
2014-07-17 18:01:08 +08:00
|
|
|
RADEON_GEM_DOMAIN_VRAM, 0, NULL,
|
2014-09-18 20:11:56 +08:00
|
|
|
NULL, &vm->page_directory);
|
2014-02-20 20:42:17 +08:00
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
r = radeon_vm_clear_bo(rdev, vm->page_directory);
|
|
|
|
if (r) {
|
|
|
|
radeon_bo_unref(&vm->page_directory);
|
|
|
|
vm->page_directory = NULL;
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* radeon_vm_fini - tear down a vm instance
|
|
|
|
*
|
|
|
|
* @rdev: radeon_device pointer
|
|
|
|
* @vm: requested vm
|
|
|
|
*
|
|
|
|
* Tear down @vm (cayman+).
|
|
|
|
* Unbind the VM and remove all bos from the vm bo list
|
|
|
|
*/
|
|
|
|
void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
|
|
|
|
{
|
|
|
|
struct radeon_bo_va *bo_va, *tmp;
|
2014-02-20 20:42:17 +08:00
|
|
|
int i, r;
|
2014-02-20 17:25:15 +08:00
|
|
|
|
2014-07-30 23:49:56 +08:00
|
|
|
if (!RB_EMPTY_ROOT(&vm->va)) {
|
2014-02-20 17:25:15 +08:00
|
|
|
dev_err(rdev->dev, "still active bo inside vm\n");
|
|
|
|
}
|
2014-07-30 23:49:56 +08:00
|
|
|
rbtree_postorder_for_each_entry_safe(bo_va, tmp, &vm->va, it.rb) {
|
|
|
|
interval_tree_remove(&bo_va->it, &vm->va);
|
2014-02-20 17:25:15 +08:00
|
|
|
r = radeon_bo_reserve(bo_va->bo, false);
|
|
|
|
if (!r) {
|
|
|
|
list_del_init(&bo_va->bo_list);
|
|
|
|
radeon_bo_unreserve(bo_va->bo);
|
2014-11-19 21:01:26 +08:00
|
|
|
radeon_fence_unref(&bo_va->last_pt_update);
|
2014-02-20 17:25:15 +08:00
|
|
|
kfree(bo_va);
|
|
|
|
}
|
|
|
|
}
|
2014-07-31 03:04:57 +08:00
|
|
|
list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
|
|
|
|
radeon_bo_unref(&bo_va->bo);
|
2014-11-19 21:01:26 +08:00
|
|
|
radeon_fence_unref(&bo_va->last_pt_update);
|
2014-07-18 14:56:40 +08:00
|
|
|
kfree(bo_va);
|
2014-07-31 03:04:57 +08:00
|
|
|
}
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
|
|
|
|
radeon_bo_unref(&vm->page_tables[i].bo);
|
|
|
|
kfree(vm->page_tables);
|
|
|
|
|
|
|
|
radeon_bo_unref(&vm->page_directory);
|
|
|
|
|
2014-11-19 21:01:25 +08:00
|
|
|
for (i = 0; i < RADEON_NUM_RINGS; ++i) {
|
|
|
|
radeon_fence_unref(&vm->ids[i].flushed_updates);
|
|
|
|
radeon_fence_unref(&vm->ids[i].last_id_use);
|
|
|
|
}
|
2014-02-20 20:42:17 +08:00
|
|
|
|
|
|
|
mutex_destroy(&vm->mutex);
|
2014-02-20 17:25:15 +08:00
|
|
|
}
|