drm/radeon: split PT setup in more functions
Move the decision what to use into the common VM code.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c
index 66325ef..8a3e622 100644
--- a/drivers/gpu/drm/radeon/ni_dma.c
+++ b/drivers/gpu/drm/radeon/ni_dma.c
@@ -307,7 +307,43 @@
}
/**
- * cayman_dma_vm_set_page - update the page tables using the DMA
+ * cayman_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 (cayman/TN).
+ */
+void cayman_dma_vm_copy_pages(struct radeon_device *rdev,
+ struct radeon_ib *ib,
+ uint64_t pe, uint64_t src,
+ unsigned count)
+{
+ unsigned ndw;
+
+ while (count) {
+ ndw = count * 2;
+ if (ndw > 0xFFFFE)
+ ndw = 0xFFFFE;
+
+ ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
+ 0, 0, ndw);
+ 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 += ndw * 4;
+ src += ndw * 4;
+ count -= ndw / 2;
+ }
+}
+
+/**
+ * cayman_dma_vm_write_pages - update PTEs by writing them manually
*
* @rdev: radeon_device pointer
* @ib: indirect buffer to fill with commands
@@ -315,90 +351,103 @@
* @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
+ * @flags: hw access flags
*
- * Update the page tables using the DMA (cayman/TN).
+ * Update PTEs by writing them manually using the DMA (cayman/TN).
*/
-void cayman_dma_vm_set_page(struct radeon_device *rdev,
- struct radeon_ib *ib,
- uint64_t pe,
- uint64_t addr, unsigned count,
- uint32_t incr, uint32_t flags)
+void cayman_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;
- trace_radeon_vm_set_page(pe, addr, count, incr, flags);
+ while (count) {
+ ndw = count * 2;
+ if (ndw > 0xFFFFE)
+ ndw = 0xFFFFE;
- if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
- uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
- while (count) {
- ndw = count * 2;
- if (ndw > 0xFFFFE)
- ndw = 0xFFFFE;
-
- ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
- 0, 0, ndw);
- 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 += ndw * 4;
- src += ndw * 4;
- count -= ndw / 2;
- }
-
- } else if ((flags & R600_PTE_SYSTEM) || (count == 1)) {
- 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, 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);
- value &= 0xFFFFFFFFFFFFF000ULL;
- } 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);
- }
- }
- } else {
- while (count) {
- ndw = count * 2;
- if (ndw > 0xFFFFE)
- ndw = 0xFFFFE;
-
- if (flags & R600_PTE_VALID)
+ /* for non-physically contiguous pages (system) */
+ ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE,
+ 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);
+ value &= 0xFFFFFFFFFFFFF000ULL;
+ } else if (flags & R600_PTE_VALID) {
value = addr;
- else
+ } 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 */
+ }
+ addr += incr;
+ value |= flags;
+ ib->ptr[ib->length_dw++] = 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;
}
}
+}
+
+/**
+ * cayman_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: hw access flags
+ *
+ * Update the page tables using the DMA (cayman/TN).
+ */
+void cayman_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;
+ }
+}
+
+/**
+ * cayman_dma_vm_pad_ib - pad the IB to the required number of dw
+ *
+ * @ib: indirect buffer to fill with padding
+ *
+ */
+void cayman_dma_vm_pad_ib(struct radeon_ib *ib)
+{
while (ib->length_dw & 0x7)
ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0);
}