Release 4.11 drivers/gpu/drm/radeon/si_dma.c
/*
* Copyright 2013 Advanced Micro Devices, Inc.
*
* 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: Alex Deucher
*/
#include <drm/drmP.h>
#include "radeon.h"
#include "radeon_asic.h"
#include "radeon_trace.h"
#include "sid.h"
u32 si_gpu_check_soft_reset(struct radeon_device *rdev);
/**
* si_dma_is_lockup - Check if the DMA engine is locked up
*
* @rdev: radeon_device pointer
* @ring: radeon_ring structure holding ring information
*
* Check if the async DMA engine is locked up.
* Returns true if the engine appears to be locked up, false if not.
*/
bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
{
u32 reset_mask = si_gpu_check_soft_reset(rdev);
u32 mask;
if (ring->idx == R600_RING_TYPE_DMA_INDEX)
mask = RADEON_RESET_DMA;
else
mask = RADEON_RESET_DMA1;
if (!(reset_mask & mask)) {
radeon_ring_lockup_update(rdev, ring);
return false;
}
return radeon_ring_test_lockup(rdev, ring);
}
Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 72 | 100.00% | 2 | 100.00% |
| Total | 72 | 100.00% | 2 | 100.00% |
/**
* si_dma_vm_copy_pages - update PTEs by copying them from the GART
*
* @rdev: radeon_device pointer
* @ib: indirect buffer to fill with commands
* @pe: addr of the page entry
* @src: src addr where to copy from
* @count: number of page entries to update
*
* Update PTEs by copying them from the GART using the DMA (SI).
*/
void si_dma_vm_copy_pages(struct radeon_device *rdev,
struct radeon_ib *ib,
uint64_t pe, uint64_t src,
unsigned count)
{
while (count) {
unsigned bytes = count * 8;
if (bytes > 0xFFFF8)
bytes = 0xFFFF8;
ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
1, 0, 0, bytes);
ib->ptr[ib->length_dw++] = lower_32_bits(pe);
ib->ptr[ib->length_dw++] = lower_32_bits(src);
ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
pe += bytes;
src += bytes;
count -= bytes / 8;
}
}
Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 148 | 100.00% | 4 | 100.00% |
| Total | 148 | 100.00% | 4 | 100.00% |
/**
* si_dma_vm_write_pages - update PTEs by writing them manually
*
* @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: access flags
*
* Update PTEs by writing them manually using the DMA (SI).
*/
void si_dma_vm_write_pages(struct radeon_device *rdev,
struct radeon_ib *ib,
uint64_t pe,
uint64_t addr, unsigned count,
uint32_t incr, uint32_t flags)
{
uint64_t value;
unsigned ndw;
while (count) {
ndw = count * 2;
if (ndw > 0xFFFFE)
ndw = 0xFFFFE;
/* for non-physically contiguous pages (system) */
ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
ib->ptr[ib->length_dw++] = pe;
ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
for (; ndw > 0; ndw -= 2, --count, pe += 8) {
if (flags & R600_PTE_SYSTEM) {
value = radeon_vm_map_gart(rdev, addr);
} else if (flags & R600_PTE_VALID) {
value = addr;
} else {
value = 0;
}
addr += incr;
value |= flags;
ib->ptr[ib->length_dw++] = value;
ib->ptr[ib->length_dw++] = upper_32_bits(value);
}
}
}
Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 203 | 100.00% | 4 | 100.00% |
| Total | 203 | 100.00% | 4 | 100.00% |
/**
* si_dma_vm_set_pages - update the page tables using the DMA
*
* @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: access flags
*
* Update the page tables using the DMA (SI).
*/
void si_dma_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)
{
uint64_t value;
unsigned ndw;
while (count) {
ndw = count * 2;
if (ndw > 0xFFFFE)
ndw = 0xFFFFE;
if (flags & R600_PTE_VALID)
value = addr;
else
value = 0;
/* for physically contiguous pages (vram) */
ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
ib->ptr[ib->length_dw++] = pe; /* dst addr */
ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
ib->ptr[ib->length_dw++] = flags; /* mask */
ib->ptr[ib->length_dw++] = 0;
ib->ptr[ib->length_dw++] = value; /* value */
ib->ptr[ib->length_dw++] = upper_32_bits(value);
ib->ptr[ib->length_dw++] = incr; /* increment size */
ib->ptr[ib->length_dw++] = 0;
pe += ndw * 4;
addr += (ndw / 2) * incr;
count -= ndw / 2;
}
}
Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 219 | 100.00% | 3 | 100.00% |
| Total | 219 | 100.00% | 3 | 100.00% |
void si_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
unsigned vm_id, uint64_t pd_addr)
{
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
if (vm_id < 8) {
radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2));
} else {
radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm_id - 8) << 2)) >> 2));
}
radeon_ring_write(ring, pd_addr >> 12);
/* flush hdp cache */
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2));
radeon_ring_write(ring, 1);
/* bits 0-7 are the VM contexts0-7 */
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2));
radeon_ring_write(ring, 1 << vm_id);
/* wait for invalidate to complete */
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_POLL_REG_MEM, 0, 0, 0, 0));
radeon_ring_write(ring, VM_INVALIDATE_REQUEST);
radeon_ring_write(ring, 0xff << 16); /* retry */
radeon_ring_write(ring, 1 << vm_id); /* mask */
radeon_ring_write(ring, 0); /* value */
radeon_ring_write(ring, (0 << 28) | 0x20); /* func(always) | poll interval */
}
Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 201 | 74.72% | 2 | 66.67% |
| Alex Deucher | 68 | 25.28% | 1 | 33.33% |
| Total | 269 | 100.00% | 3 | 100.00% |
/**
* si_copy_dma - copy pages using the DMA engine
*
* @rdev: radeon_device pointer
* @src_offset: src GPU address
* @dst_offset: dst GPU address
* @num_gpu_pages: number of GPU pages to xfer
* @resv: reservation object to sync to
*
* Copy GPU paging using the DMA engine (SI).
* Used by the radeon ttm implementation to move pages if
* registered as the asic copy callback.
*/
struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
uint64_t src_offset, uint64_t dst_offset,
unsigned num_gpu_pages,
struct reservation_object *resv)
{
struct radeon_fence *fence;
struct radeon_sync sync;
int ring_index = rdev->asic->copy.dma_ring_index;
struct radeon_ring *ring = &rdev->ring[ring_index];
u32 size_in_bytes, cur_size_in_bytes;
int i, num_loops;
int r = 0;
radeon_sync_create(&sync);
size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
num_loops = DIV_ROUND_UP(size_in_bytes, 0xfffff);
r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11);
if (r) {
DRM_ERROR("radeon: moving bo (%d).\n", r);
radeon_sync_free(rdev, &sync, NULL);
return ERR_PTR(r);
}
radeon_sync_resv(rdev, &sync, resv, false);
radeon_sync_rings(rdev, &sync, ring->idx);
for (i = 0; i < num_loops; i++) {
cur_size_in_bytes = size_in_bytes;
if (cur_size_in_bytes > 0xFFFFF)
cur_size_in_bytes = 0xFFFFF;
size_in_bytes -= cur_size_in_bytes;
radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 1, 0, 0, cur_size_in_bytes));
radeon_ring_write(ring, lower_32_bits(dst_offset));
radeon_ring_write(ring, lower_32_bits(src_offset));
radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff);
radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff);
src_offset += cur_size_in_bytes;
dst_offset += cur_size_in_bytes;
}
r = radeon_fence_emit(rdev, &fence, ring->idx);
if (r) {
radeon_ring_unlock_undo(rdev, ring);
radeon_sync_free(rdev, &sync, NULL);
return ERR_PTR(r);
}
radeon_ring_unlock_commit(rdev, ring, false);
radeon_sync_free(rdev, &sync, fence);
return fence;
}
Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 321 | 96.40% | 5 | 62.50% |
| Maarten Lankhorst | 10 | 3.00% | 2 | 25.00% |
| Michel Dänzer | 2 | 0.60% | 1 | 12.50% |
| Total | 333 | 100.00% | 8 | 100.00% |
Overall Contributors
| Person | Tokens | Prop | Commits | CommitProp |
| Christian König | 1194 | 93.72% | 11 | 73.33% |
| Alex Deucher | 68 | 5.34% | 1 | 6.67% |
| Maarten Lankhorst | 10 | 0.78% | 2 | 13.33% |
| Michel Dänzer | 2 | 0.16% | 1 | 6.67% |
| Total | 1274 | 100.00% | 15 | 100.00% |
Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.