Commit ac01db3d authored by Jammy Zhou's avatar Jammy Zhou Committed by Alex Deucher

drm/amdgpu: implement burst NOP for SDMA

Customize the insert_nop func for SDMA rings, and use burst NOP for
ring/IB submissions in other places as well
Signed-off-by: default avatarJammy Zhou <Jammy.Zhou@amd.com>
Reviewed-by: default avatarAlex Deucher <alexander.deucher@amd.com>
parent edff0e28
......@@ -188,6 +188,19 @@ static void cik_sdma_ring_set_wptr(struct amdgpu_ring *ring)
WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[me], (ring->wptr << 2) & 0x3fffc);
}
static void cik_sdma_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
{
struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ring);
int i;
for (i = 0; i < count; i++)
if (sdma && sdma->burst_nop && (i == 0))
amdgpu_ring_write(ring, ring->nop |
SDMA_NOP_COUNT(count - 1));
else
amdgpu_ring_write(ring, ring->nop);
}
/**
* cik_sdma_ring_emit_ib - Schedule an IB on the DMA engine
*
......@@ -213,8 +226,8 @@ static void cik_sdma_ring_emit_ib(struct amdgpu_ring *ring,
amdgpu_ring_write(ring, next_rptr);
/* IB packet must end on a 8 DW boundary */
while ((ring->wptr & 7) != 4)
amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
cik_sdma_ring_insert_nop(ring, (12 - (ring->wptr & 7)) % 8);
amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_INDIRECT_BUFFER, 0, extra_bits));
amdgpu_ring_write(ring, ib->gpu_addr & 0xffffffe0); /* base must be 32 byte aligned */
amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xffffffff);
......@@ -817,8 +830,19 @@ static void cik_sdma_vm_set_pte_pde(struct amdgpu_ib *ib,
*/
static void cik_sdma_vm_pad_ib(struct amdgpu_ib *ib)
{
while (ib->length_dw & 0x7)
ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0);
struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ib->ring);
u32 pad_count;
int i;
pad_count = (8 - (ib->length_dw & 0x7)) % 8;
for (i = 0; i < pad_count; i++)
if (sdma && sdma->burst_nop && (i == 0))
ib->ptr[ib->length_dw++] =
SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0) |
SDMA_NOP_COUNT(pad_count - 1);
else
ib->ptr[ib->length_dw++] =
SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0);
}
/**
......@@ -1305,7 +1329,7 @@ static const struct amdgpu_ring_funcs cik_sdma_ring_funcs = {
.test_ring = cik_sdma_ring_test_ring,
.test_ib = cik_sdma_ring_test_ib,
.is_lockup = cik_sdma_ring_is_lockup,
.insert_nop = amdgpu_ring_insert_nop,
.insert_nop = cik_sdma_ring_insert_nop,
};
static void cik_sdma_set_ring_funcs(struct amdgpu_device *adev)
......
......@@ -220,6 +220,19 @@ static void sdma_v2_4_ring_set_wptr(struct amdgpu_ring *ring)
WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[me], ring->wptr << 2);
}
static void sdma_v2_4_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
{
struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ring);
int i;
for (i = 0; i < count; i++)
if (sdma && sdma->burst_nop && (i == 0))
amdgpu_ring_write(ring, ring->nop |
SDMA_PKT_NOP_HEADER_COUNT(count - 1));
else
amdgpu_ring_write(ring, ring->nop);
}
/**
* sdma_v2_4_ring_emit_ib - Schedule an IB on the DMA engine
*
......@@ -247,8 +260,8 @@ static void sdma_v2_4_ring_emit_ib(struct amdgpu_ring *ring,
amdgpu_ring_write(ring, next_rptr);
/* IB packet must end on a 8 DW boundary */
while ((ring->wptr & 7) != 2)
amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_NOP));
sdma_v2_4_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8);
amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) |
SDMA_PKT_INDIRECT_HEADER_VMID(vmid));
/* base must be 32 byte aligned */
......@@ -881,8 +894,19 @@ static void sdma_v2_4_vm_set_pte_pde(struct amdgpu_ib *ib,
*/
static void sdma_v2_4_vm_pad_ib(struct amdgpu_ib *ib)
{
while (ib->length_dw & 0x7)
ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_NOP);
struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ib->ring);
u32 pad_count;
int i;
pad_count = (8 - (ib->length_dw & 0x7)) % 8;
for (i = 0; i < pad_count; i++)
if (sdma && sdma->burst_nop && (i == 0))
ib->ptr[ib->length_dw++] =
SDMA_PKT_HEADER_OP(SDMA_OP_NOP) |
SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1);
else
ib->ptr[ib->length_dw++] =
SDMA_PKT_HEADER_OP(SDMA_OP_NOP);
}
/**
......@@ -1316,7 +1340,7 @@ static const struct amdgpu_ring_funcs sdma_v2_4_ring_funcs = {
.test_ring = sdma_v2_4_ring_test_ring,
.test_ib = sdma_v2_4_ring_test_ib,
.is_lockup = sdma_v2_4_ring_is_lockup,
.insert_nop = amdgpu_ring_insert_nop,
.insert_nop = sdma_v2_4_ring_insert_nop,
};
static void sdma_v2_4_set_ring_funcs(struct amdgpu_device *adev)
......
......@@ -306,6 +306,19 @@ static void sdma_v3_0_ring_set_wptr(struct amdgpu_ring *ring)
}
}
static void sdma_v3_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
{
struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ring);
int i;
for (i = 0; i < count; i++)
if (sdma && sdma->burst_nop && (i == 0))
amdgpu_ring_write(ring, ring->nop |
SDMA_PKT_NOP_HEADER_COUNT(count - 1));
else
amdgpu_ring_write(ring, ring->nop);
}
/**
* sdma_v3_0_ring_emit_ib - Schedule an IB on the DMA engine
*
......@@ -332,8 +345,7 @@ static void sdma_v3_0_ring_emit_ib(struct amdgpu_ring *ring,
amdgpu_ring_write(ring, next_rptr);
/* IB packet must end on a 8 DW boundary */
while ((ring->wptr & 7) != 2)
amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_NOP));
sdma_v3_0_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8);
amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_INDIRECT) |
SDMA_PKT_INDIRECT_HEADER_VMID(vmid));
......@@ -1001,8 +1013,19 @@ static void sdma_v3_0_vm_set_pte_pde(struct amdgpu_ib *ib,
*/
static void sdma_v3_0_vm_pad_ib(struct amdgpu_ib *ib)
{
while (ib->length_dw & 0x7)
ib->ptr[ib->length_dw++] = SDMA_PKT_HEADER_OP(SDMA_OP_NOP);
struct amdgpu_sdma *sdma = amdgpu_get_sdma_instance(ib->ring);
u32 pad_count;
int i;
pad_count = (8 - (ib->length_dw & 0x7)) % 8;
for (i = 0; i < pad_count; i++)
if (sdma && sdma->burst_nop && (i == 0))
ib->ptr[ib->length_dw++] =
SDMA_PKT_HEADER_OP(SDMA_OP_NOP) |
SDMA_PKT_NOP_HEADER_COUNT(pad_count - 1);
else
ib->ptr[ib->length_dw++] =
SDMA_PKT_HEADER_OP(SDMA_OP_NOP);
}
/**
......@@ -1440,7 +1463,7 @@ static const struct amdgpu_ring_funcs sdma_v3_0_ring_funcs = {
.test_ring = sdma_v3_0_ring_test_ring,
.test_ib = sdma_v3_0_ring_test_ib,
.is_lockup = sdma_v3_0_ring_is_lockup,
.insert_nop = amdgpu_ring_insert_nop,
.insert_nop = sdma_v3_0_ring_insert_nop,
};
static void sdma_v3_0_set_ring_funcs(struct amdgpu_device *adev)
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment