/**
* cik_sdma_ib_test - test an IB on the DMA engine
*
* @rdev: radeon_device pointer
* @ring: radeon_ring structure holding ring information
*
* Test a simple IB in the DMA ring (CIK).
* Returns 0 on success, error on failure.
*/
int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
{
struct radeon_ib ib;
unsigned i;
unsigned index;
int r;
u32 tmp = 0;
u64 gpu_addr;
if (ring->idx == R600_RING_TYPE_DMA_INDEX)
index = R600_WB_DMA_RING_TEST_OFFSET;
else
index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
gpu_addr = rdev->wb.gpu_addr + index;
tmp = 0xCAFEDEAD;
rdev->wb.wb[index/4] = cpu_to_le32(tmp);
r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
if (r) {
DRM_ERROR("radeon: failed to get ib (%d).\n", r);
return r;
}
ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
ib.ptr[1] = lower_32_bits(gpu_addr);
ib.ptr[2] = upper_32_bits(gpu_addr);
ib.ptr[3] = 1;
ib.ptr[4] = 0xDEADBEEF;
ib.length_dw = 5;
r = radeon_ib_schedule(rdev, &ib, NULL, false);
if (r) {
radeon_ib_free(rdev, &ib);
DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
return r;
}
r = radeon_fence_wait(ib.fence, false);
if (r) {
DRM_ERROR("radeon: fence wait failed (%d).\n", r);
return r;
}
for (i = 0; i < rdev->usec_timeout; i++) {
tmp = le32_to_cpu(rdev->wb.wb[index/4]);
if (tmp == 0xDEADBEEF)
break;
DRM_UDELAY(1);
}
if (i < rdev->usec_timeout) {
DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i);
} else {
DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp);
r = -EINVAL;
}
radeon_ib_free(rdev, &ib);
return r;
}
/**
* cik_sdma_ib_test - test an IB on the DMA engine
*
* @rdev: radeon_device pointer
* @ring: radeon_ring structure holding ring information
*
* Test a simple IB in the DMA ring (CIK).
* Returns 0 on success, error on failure.
*/
int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
{
struct radeon_ib ib;
unsigned i;
int r;
void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
u32 tmp = 0;
if (!ptr) {
DRM_ERROR("invalid vram scratch pointer\n");
return -EINVAL;
}
tmp = 0xCAFEDEAD;
writel(tmp, ptr);
r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
if (r) {
DRM_ERROR("radeon: failed to get ib (%d).\n", r);
return r;
}
ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr);
ib.ptr[3] = 1;
ib.ptr[4] = 0xDEADBEEF;
ib.length_dw = 5;
r = radeon_ib_schedule(rdev, &ib, NULL);
if (r) {
radeon_ib_free(rdev, &ib);
DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
return r;
}
r = radeon_fence_wait(ib.fence, false);
if (r) {
DRM_ERROR("radeon: fence wait failed (%d).\n", r);
return r;
}
for (i = 0; i < rdev->usec_timeout; i++) {
tmp = readl(ptr);
if (tmp == 0xDEADBEEF)
break;
DRM_UDELAY(1);
}
if (i < rdev->usec_timeout) {
DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i);
} else {
DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp);
r = -EINVAL;
}
radeon_ib_free(rdev, &ib);
return r;
}
/**
* cs_parser_fini() - clean parser states
* @parser: parser structure holding parsing context.
* @error: error number
*
* If error is set than unvalidate buffer, otherwise just free memory
* used by parsing context.
**/
static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
{
unsigned i;
if (!error && parser->ib)
ttm_eu_fence_buffer_objects(&parser->validated,
parser->ib->fence);
else
ttm_eu_backoff_reservation(&parser->validated);
if (parser->relocs != NULL) {
for (i = 0; i < parser->nrelocs; i++) {
if (parser->relocs[i].gobj)
drm_gem_object_unreference_unlocked(parser->relocs[i].gobj);
}
}
kfree(parser->track);
kfree(parser->relocs);
kfree(parser->relocs_ptr);
for (i = 0; i < parser->nchunks; i++) {
kfree(parser->chunks[i].kdata);
kfree(parser->chunks[i].kpage[0]);
kfree(parser->chunks[i].kpage[1]);
}
kfree(parser->chunks);
kfree(parser->chunks_array);
radeon_ib_free(parser->rdev, &parser->ib);
}
/**
* cs_parser_fini() - clean parser states
* @parser: parser structure holding parsing context.
* @error: error number
*
* If error is set than unvalidate buffer, otherwise just free memory
* used by parsing context.
**/
static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
{
unsigned i;
if (error) {
radeon_object_list_unvalidate(&parser->validated);
} else {
radeon_object_list_clean(&parser->validated);
}
for (i = 0; i < parser->nrelocs; i++) {
if (parser->relocs[i].gobj) {
mutex_lock(&parser->rdev->ddev->struct_mutex);
drm_gem_object_unreference(parser->relocs[i].gobj);
mutex_unlock(&parser->rdev->ddev->struct_mutex);
}
}
kfree(parser->track);
kfree(parser->relocs);
kfree(parser->relocs_ptr);
for (i = 0; i < parser->nchunks; i++) {
kfree(parser->chunks[i].kdata);
kfree(parser->chunks[i].kpage[0]);
kfree(parser->chunks[i].kpage[1]);
}
kfree(parser->chunks);
kfree(parser->chunks_array);
radeon_ib_free(parser->rdev, &parser->ib);
}
/**
* radeon_vce_get_create_msg - generate a VCE create msg
*
* @rdev: radeon_device pointer
* @ring: ring we should submit the msg to
* @handle: VCE session handle to use
* @fence: optional fence to return
*
* Open up a stream for HW test
*/
int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring,
uint32_t handle, struct radeon_fence **fence)
{
const unsigned ib_size_dw = 1024;
struct radeon_ib ib;
uint64_t dummy;
int i, r;
r = radeon_ib_get(rdev, ring, &ib, NULL, ib_size_dw * 4);
if (r) {
DRM_ERROR("radeon: failed to get ib (%d).\n", r);
return r;
}
dummy = ib.gpu_addr + 1024;
/* stitch together an VCE create msg */
ib.length_dw = 0;
ib.ptr[ib.length_dw++] = 0x0000000c; /* len */
ib.ptr[ib.length_dw++] = 0x00000001; /* session cmd */
ib.ptr[ib.length_dw++] = handle;
ib.ptr[ib.length_dw++] = 0x00000030; /* len */
ib.ptr[ib.length_dw++] = 0x01000001; /* create cmd */
ib.ptr[ib.length_dw++] = 0x00000000;
ib.ptr[ib.length_dw++] = 0x00000042;
ib.ptr[ib.length_dw++] = 0x0000000a;
ib.ptr[ib.length_dw++] = 0x00000001;
ib.ptr[ib.length_dw++] = 0x00000080;
ib.ptr[ib.length_dw++] = 0x00000060;
ib.ptr[ib.length_dw++] = 0x00000100;
ib.ptr[ib.length_dw++] = 0x00000100;
ib.ptr[ib.length_dw++] = 0x0000000c;
ib.ptr[ib.length_dw++] = 0x00000000;
ib.ptr[ib.length_dw++] = 0x00000014; /* len */
ib.ptr[ib.length_dw++] = 0x05000005; /* feedback buffer */
ib.ptr[ib.length_dw++] = upper_32_bits(dummy);
ib.ptr[ib.length_dw++] = dummy;
ib.ptr[ib.length_dw++] = 0x00000001;
for (i = ib.length_dw; i < ib_size_dw; ++i)
ib.ptr[i] = 0x0;
r = radeon_ib_schedule(rdev, &ib, NULL);
if (r) {
DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
}
if (fence)
*fence = radeon_fence_ref(ib.fence);
radeon_ib_free(rdev, &ib);
return r;
}
/**
* cs_parser_fini() - clean parser states
* @parser: parser structure holding parsing context.
* @error: error number
*
* If error is set than unvalidate buffer, otherwise just free memory
* used by parsing context.
**/
static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bool backoff)
{
unsigned i;
if (!error) {
/* Sort the buffer list from the smallest to largest buffer,
* which affects the order of buffers in the LRU list.
* This assures that the smallest buffers are added first
* to the LRU list, so they are likely to be later evicted
* first, instead of large buffers whose eviction is more
* expensive.
*
* This slightly lowers the number of bytes moved by TTM
* per frame under memory pressure.
*/
list_sort(NULL, &parser->validated, cmp_size_smaller_first);
ttm_eu_fence_buffer_objects(&parser->ticket,
&parser->validated,
parser->ib.fence);
} else if (backoff) {
ttm_eu_backoff_reservation(&parser->ticket,
&parser->validated);
}
if (parser->relocs != NULL) {
for (i = 0; i < parser->nrelocs; i++) {
if (parser->relocs[i].gobj)
drm_gem_object_unreference_unlocked(parser->relocs[i].gobj);
}
}
kfree(parser->track);
kfree(parser->relocs);
kfree(parser->relocs_ptr);
kfree(parser->vm_bos);
for (i = 0; i < parser->nchunks; i++)
drm_free_large(parser->chunks[i].kdata);
kfree(parser->chunks);
kfree(parser->chunks_array);
radeon_ib_free(parser->rdev, &parser->ib);
radeon_ib_free(parser->rdev, &parser->const_ib);
}
/**
* radeon_vce_get_destroy_msg - generate a VCE destroy msg
*
* @rdev: radeon_device pointer
* @ring: ring we should submit the msg to
* @handle: VCE session handle to use
* @fence: optional fence to return
*
* Close up a stream for HW test or if userspace failed to do so
*/
int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring,
uint32_t handle, struct radeon_fence **fence)
{
const unsigned ib_size_dw = 1024;
struct radeon_ib ib;
uint64_t dummy;
int i, r;
r = radeon_ib_get(rdev, ring, &ib, NULL, ib_size_dw * 4);
if (r) {
DRM_ERROR("radeon: failed to get ib (%d).\n", r);
return r;
}
dummy = ib.gpu_addr + 1024;
/* stitch together an VCE destroy msg */
ib.length_dw = 0;
ib.ptr[ib.length_dw++] = cpu_to_le32(0x0000000c); /* len */
ib.ptr[ib.length_dw++] = cpu_to_le32(0x00000001); /* session cmd */
ib.ptr[ib.length_dw++] = cpu_to_le32(handle);
ib.ptr[ib.length_dw++] = cpu_to_le32(0x00000014); /* len */
ib.ptr[ib.length_dw++] = cpu_to_le32(0x05000005); /* feedback buffer */
ib.ptr[ib.length_dw++] = cpu_to_le32(upper_32_bits(dummy));
ib.ptr[ib.length_dw++] = cpu_to_le32(dummy);
ib.ptr[ib.length_dw++] = cpu_to_le32(0x00000001);
ib.ptr[ib.length_dw++] = cpu_to_le32(0x00000008); /* len */
ib.ptr[ib.length_dw++] = cpu_to_le32(0x02000001); /* destroy cmd */
for (i = ib.length_dw; i < ib_size_dw; ++i)
ib.ptr[i] = cpu_to_le32(0x0);
r = radeon_ib_schedule(rdev, &ib, NULL, false);
if (r) {
DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
}
if (fence)
*fence = radeon_fence_ref(ib.fence);
radeon_ib_free(rdev, &ib);
return r;
}