AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

r600g: rework queries

Browse Source 
We always mapped the query buffer in begin_query, causing stalls
if the buffer was busy.

This commit reworks it such that the query buffer is only mapped
in get_query_result as it's supposed to be.

The query buffer is no longer treated as a ring buffer. Instead, the results
are just appended and when the buffer is full, we create a new one. One query
can have more than one query buffer, though that's a very rare case.
Begin_query releases all query buffers.

Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Marek Olšák
2012-02-22 00:25:55 +01:00
parent 5b84a8c3c3
commit df00dc3c81
3 changed files with 167 additions and 116 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/gallium/drivers/r600/r600.h
+21 -14
View File
@@ -157,25 +157,31 @@ struct r600_range {
struct r600_block **blocks;
};
struct r600_query {
union {
uint64_t u64;
boolean b;
struct pipe_query_data_so_statistics so;
} result;
/* The kind of query */
unsigned type;
/* Offset of the first result for current query */
unsigned results_start;
struct r600_query_buffer {
/* The buffer where query results are stored. */
struct r600_resource *buf;
/* Offset of the next free result after current query data */
unsigned results_end;
/* If a query buffer is full, a new buffer is created and the old one
* is put in here. When we calculate the result, we sum up the samples
* from all buffers. */
struct r600_query_buffer *previous;
};
union r600_query_result {
uint64_t u64;
boolean b;
struct pipe_query_data_so_statistics so;
};
struct r600_query {
/* The query buffer and how many results are in it. */
struct r600_query_buffer buffer;
/* The type of query */
unsigned type;
/* Size of the result in memory for both begin_query and end_query,
* this can be one or two numbers, or it could even be a size of a structure. */
unsigned result_size;
/* The buffer where query results are stored. It's used as a ring,
* data blocks for current query are stored sequentially from
* results_start to results_end, with wrapping on the buffer end */
struct r600_resource *buffer;
/* The number of dwords for begin_query or end_query. */
unsigned num_cs_dw;
/* linked list of queries */
@@ -214,6 +220,7 @@ void r600_context_query_destroy(struct r600_context *ctx, struct r600_query *que
boolean r600_context_query_result(struct r600_context *ctx,
struct r600_query *query,
boolean wait, void *vresult);
struct r600_resource *r600_new_query_buffer(struct r600_context *ctx, unsigned type);
void r600_query_begin(struct r600_context *ctx, struct r600_query *query);
void r600_query_end(struct r600_context *ctx, struct r600_query *query);
void r600_context_queries_suspend(struct r600_context *ctx);
src/gallium/drivers/r600/r600_hw_context.c
+119 -91
View File
@@ -1349,12 +1349,16 @@ static unsigned r600_query_read_result(char *map, unsigned start_index, unsigned
return 0;
}
static boolean r600_query_result(struct r600_context *ctx, struct r600_query *query, boolean wait)
static boolean r600_query_buffer_result(struct r600_context *ctx,
struct r600_query *query,
struct r600_query_buffer *qbuf,
boolean wait,
union r600_query_result *result)
{
unsigned results_base = query->results_start;
unsigned results_base = 0;
char *map;
map = ctx->ws->buffer_map(query->buffer->buf, ctx->cs,
map = ctx->ws->buffer_map(qbuf->buf->buf, ctx->cs,
PIPE_TRANSFER_READ |
(wait ? 0 : PIPE_TRANSFER_DONTBLOCK));
if (!map)
@@ -1363,24 +1367,24 @@ static boolean r600_query_result(struct r600_context *ctx, struct r600_query *qu
/* count all results across all data blocks */
switch (query->type) {
case PIPE_QUERY_OCCLUSION_COUNTER:
while (results_base != query->results_end) {
query->result.u64 +=
while (results_base != qbuf->results_end) {
result->u64 +=
r600_query_read_result(map + results_base, 0, 2, true);
results_base = (results_base + 16) % query->buffer->b.b.b.width0;
results_base += 16;
}
break;
case PIPE_QUERY_OCCLUSION_PREDICATE:
while (results_base != query->results_end) {
query->result.b = query->result.b ||
while (results_base != qbuf->results_end) {
result->b = result->b ||
r600_query_read_result(map + results_base, 0, 2, true) != 0;
results_base = (results_base + 16) % query->buffer->b.b.b.width0;
results_base += 16;
}
break;
case PIPE_QUERY_TIME_ELAPSED:
while (results_base != query->results_end) {
query->result.u64 +=
while (results_base != qbuf->results_end) {
result->u64 +=
r600_query_read_result(map + results_base, 0, 2, false);
results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
results_base += query->result_size;
}
break;
case PIPE_QUERY_PRIMITIVES_EMITTED:
@@ -1390,79 +1394,84 @@ static boolean r600_query_result(struct r600_context *ctx, struct r600_query *qu
* u64 PrimitiveStorageNeeded;
* }
* We only need NumPrimitivesWritten here. */
while (results_base != query->results_end) {
query->result.u64 +=
while (results_base != qbuf->results_end) {
result->u64 +=
r600_query_read_result(map + results_base, 2, 6, true);
results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
results_base += query->result_size;
}
break;
case PIPE_QUERY_PRIMITIVES_GENERATED:
/* Here we read PrimitiveStorageNeeded. */
while (results_base != query->results_end) {
query->result.u64 +=
while (results_base != qbuf->results_end) {
result->u64 +=
r600_query_read_result(map + results_base, 0, 4, true);
results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
results_base += query->result_size;
}
break;
case PIPE_QUERY_SO_STATISTICS:
while (results_base != query->results_end) {
query->result.so.num_primitives_written +=
while (results_base != qbuf->results_end) {
result->so.num_primitives_written +=
r600_query_read_result(map + results_base, 2, 6, true);
query->result.so.primitives_storage_needed +=
result->so.primitives_storage_needed +=
r600_query_read_result(map + results_base, 0, 4, true);
results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
results_base += query->result_size;
}
break;
case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
while (results_base != query->results_end) {
query->result.b = query->result.b ||
while (results_base != qbuf->results_end) {
result->b = result->b ||
r600_query_read_result(map + results_base, 2, 6, true) !=
r600_query_read_result(map + results_base, 0, 4, true);
results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
results_base += query->result_size;
}
break;
default:
assert(0);
}
query->results_start = query->results_end;
ctx->ws->buffer_unmap(query->buffer->buf);
ctx->ws->buffer_unmap(qbuf->buf->buf);
return TRUE;
}
void r600_query_begin(struct r600_context *ctx, struct r600_query *query)
static boolean r600_query_result(struct r600_context *ctx, struct r600_query *query, boolean wait,
union r600_query_result *result)
{
struct radeon_winsys_cs *cs = ctx->cs;
unsigned new_results_end, i;
uint32_t *results;
uint64_t va;
struct r600_query_buffer *qbuf;
r600_need_cs_space(ctx, query->num_cs_dw * 2, TRUE);
new_results_end = (query->results_end + query->result_size) % query->buffer->b.b.b.width0;
/* collect current results if query buffer is full */
if (new_results_end == query->results_start) {
r600_query_result(ctx, query, TRUE);
for (qbuf = &query->buffer; qbuf; qbuf = qbuf->previous) {
if (!r600_query_buffer_result(ctx, query, qbuf, wait, result)) {
return FALSE;
}
}
return TRUE;
}
switch (query->type) {
struct r600_resource *r600_new_query_buffer(struct r600_context *ctx, unsigned type)
{
unsigned j, i, num_results, buf_size = 4096;
uint32_t *results;
struct r600_resource *buf = (struct r600_resource*)
pipe_buffer_create(&ctx->screen->screen, PIPE_BIND_CUSTOM,
PIPE_USAGE_STAGING, buf_size);
switch (type) {
case PIPE_QUERY_OCCLUSION_COUNTER:
case PIPE_QUERY_OCCLUSION_PREDICATE:
results = ctx->ws->buffer_map(query->buffer->buf, ctx->cs, PIPE_TRANSFER_WRITE);
if (results) {
results = (uint32_t*)((char*)results + query->results_end);
memset(results, 0, query->result_size);
results = ctx->ws->buffer_map(buf->buf, ctx->cs, PIPE_TRANSFER_WRITE);
memset(results, 0, buf_size);
/* Set top bits for unused backends */
/* Set top bits for unused backends. */
num_results = buf_size / (16 * ctx->max_db);
for (j = 0; j < num_results; j++) {
for (i = 0; i < ctx->max_db; i++) {
if (!(ctx->backend_mask & (1<<i))) {
results[(i * 4)+1] = 0x80000000;
results[(i * 4)+3] = 0x80000000;
}
}
ctx->ws->buffer_unmap(query->buffer->buf);
results += 4 * ctx->max_db;
}
ctx->ws->buffer_unmap(buf->buf);
break;
case PIPE_QUERY_TIME_ELAPSED:
break;
@@ -1470,18 +1479,35 @@ void r600_query_begin(struct r600_context *ctx, struct r600_query *query)
case PIPE_QUERY_PRIMITIVES_GENERATED:
case PIPE_QUERY_SO_STATISTICS:
case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
results = ctx->ws->buffer_map(query->buffer->buf, ctx->cs, PIPE_TRANSFER_WRITE);
results = (uint32_t*)((char*)results + query->results_end);
memset(results, 0, query->result_size);
ctx->ws->buffer_unmap(query->buffer->buf);
results = ctx->ws->buffer_map(buf->buf, ctx->cs, PIPE_TRANSFER_WRITE);
memset(results, 0, buf_size);
ctx->ws->buffer_unmap(buf->buf);
break;
default:
assert(0);
}
return buf;
}
void r600_query_begin(struct r600_context *ctx, struct r600_query *query)
{
struct radeon_winsys_cs *cs = ctx->cs;
uint64_t va;
r600_need_cs_space(ctx, query->num_cs_dw * 2, TRUE);
/* Get a new query buffer if needed. */
if (query->buffer.results_end + query->result_size > query->buffer.buf->b.b.b.width0) {
struct r600_query_buffer *qbuf = MALLOC_STRUCT(r600_query_buffer);
*qbuf = query->buffer;
query->buffer.buf = r600_new_query_buffer(ctx, query->type);
query->buffer.results_end = 0;
query->buffer.previous = qbuf;
}
/* emit begin query */
va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer);
va += query->results_end;
va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer.buf);
va += query->buffer.results_end;
switch (query->type) {
case PIPE_QUERY_OCCLUSION_COUNTER:
@@ -1497,8 +1523,8 @@ void r600_query_begin(struct r600_context *ctx, struct r600_query *query)
case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_SAMPLE_STREAMOUTSTATS) | EVENT_INDEX(3);
cs->buf[cs->cdw++] = query->results_end;
cs->buf[cs->cdw++] = 0;
cs->buf[cs->cdw++] = va;
cs->buf[cs->cdw++] = (va >> 32UL) & 0xFF;
break;
case PIPE_QUERY_TIME_ELAPSED:
cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE_EOP, 4, 0);
@@ -1512,7 +1538,7 @@ void r600_query_begin(struct r600_context *ctx, struct r600_query *query)
assert(0);
}
cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer, RADEON_USAGE_WRITE);
cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer.buf, RADEON_USAGE_WRITE);
ctx->num_cs_dw_queries_suspend += query->num_cs_dw;
}
@@ -1522,12 +1548,12 @@ void r600_query_end(struct r600_context *ctx, struct r600_query *query)
struct radeon_winsys_cs *cs = ctx->cs;
uint64_t va;
va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer);
va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer.buf);
/* emit end query */
switch (query->type) {
case PIPE_QUERY_OCCLUSION_COUNTER:
case PIPE_QUERY_OCCLUSION_PREDICATE:
va += query->results_end + 8;
va += query->buffer.results_end + 8;
cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_ZPASS_DONE) | EVENT_INDEX(1);
cs->buf[cs->cdw++] = va;
@@ -1539,11 +1565,11 @@ void r600_query_end(struct r600_context *ctx, struct r600_query *query)
case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 2, 0);
cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_SAMPLE_STREAMOUTSTATS) | EVENT_INDEX(3);
cs->buf[cs->cdw++] = query->results_end + query->result_size/2;
cs->buf[cs->cdw++] = query->buffer.results_end + query->result_size/2;
cs->buf[cs->cdw++] = 0;
break;
case PIPE_QUERY_TIME_ELAPSED:
va += query->results_end + query->result_size/2;
va += query->buffer.results_end + query->result_size/2;
cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE_EOP, 4, 0);
cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5);
cs->buf[cs->cdw++] = va;
@@ -1555,9 +1581,9 @@ void r600_query_end(struct r600_context *ctx, struct r600_query *query)
assert(0);
}
cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer, RADEON_USAGE_WRITE);
cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer.buf, RADEON_USAGE_WRITE);
query->results_end = (query->results_end + query->result_size) % query->buffer->b.b.b.width0;
query->buffer.results_end += query->result_size;
ctx->num_cs_dw_queries_suspend -= query->num_cs_dw;
}
@@ -1565,7 +1591,6 @@ void r600_query_predication(struct r600_context *ctx, struct r600_query *query,
int flag_wait)
{
struct radeon_winsys_cs *cs = ctx->cs;
uint64_t va;
if (operation == PREDICATION_OP_CLEAR) {
r600_need_cs_space(ctx, 3, FALSE);
@@ -1574,40 +1599,44 @@ void r600_query_predication(struct r600_context *ctx, struct r600_query *query,
cs->buf[cs->cdw++] = 0;
cs->buf[cs->cdw++] = PRED_OP(PREDICATION_OP_CLEAR);
} else {
unsigned results_base = query->results_start;
struct r600_query_buffer *qbuf;
unsigned count;
uint32_t op;
/* find count of the query data blocks */
count = (query->buffer->b.b.b.width0 + query->results_end - query->results_start) % query->buffer->b.b.b.width0;
count /= query->result_size;
/* Find how many results there are. */
count = 0;
for (qbuf = &query->buffer; qbuf; qbuf = qbuf->previous) {
count += qbuf->results_end / query->result_size;
}
r600_need_cs_space(ctx, 5 * count, TRUE);
op = PRED_OP(operation) | PREDICATION_DRAW_VISIBLE |
(flag_wait ? PREDICATION_HINT_WAIT : PREDICATION_HINT_NOWAIT_DRAW);
va = r600_resource_va(&ctx->screen->screen, (void*)query->buffer);
/* emit predicate packets for all data blocks */
while (results_base != query->results_end) {
cs->buf[cs->cdw++] = PKT3(PKT3_SET_PREDICATION, 1, 0);
cs->buf[cs->cdw++] = (va + results_base) & 0xFFFFFFFFUL;
cs->buf[cs->cdw++] = op | (((va + results_base) >> 32UL) & 0xFF);
cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, query->buffer,
RADEON_USAGE_READ);
results_base = (results_base + query->result_size) % query->buffer->b.b.b.width0;
for (qbuf = &query->buffer; qbuf; qbuf = qbuf->previous) {
unsigned results_base = 0;
uint64_t va = r600_resource_va(&ctx->screen->screen, &qbuf->buf->b.b.b);
/* set CONTINUE bit for all packets except the first */
op |= PREDICATION_CONTINUE;
}
while (results_base < qbuf->results_end) {
cs->buf[cs->cdw++] = PKT3(PKT3_SET_PREDICATION, 1, 0);
cs->buf[cs->cdw++] = (va + results_base) & 0xFFFFFFFFUL;
cs->buf[cs->cdw++] = op | (((va + results_base) >> 32UL) & 0xFF);
cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
cs->buf[cs->cdw++] = r600_context_bo_reloc(ctx, qbuf->buf, RADEON_USAGE_READ);
results_base += query->result_size;
/* set CONTINUE bit for all packets except the first */
op |= PREDICATION_CONTINUE;
}
} while (qbuf);
}
}
struct r600_query *r600_context_query_create(struct r600_context *ctx, unsigned query_type)
{
struct r600_query *query;
unsigned buffer_size = 4096;
query = CALLOC_STRUCT(r600_query);
if (query == NULL)
@@ -1639,16 +1668,12 @@ struct r600_query *r600_context_query_create(struct r600_context *ctx, unsigned
return NULL;
}
/* adjust buffer size to simplify offsets wrapping math */
buffer_size -= buffer_size % query->result_size;
/* Queries are normally read by the CPU after
* being written by the gpu, hence staging is probably a good
* usage pattern.
*/
query->buffer = (struct r600_resource*)
pipe_buffer_create(&ctx->screen->screen, PIPE_BIND_CUSTOM, PIPE_USAGE_STAGING, buffer_size);
if (!query->buffer) {
query->buffer.buf = r600_new_query_buffer(ctx, query_type);
if (!query->buffer.buf) {
FREE(query);
return NULL;
}
@@ -1657,7 +1682,7 @@ struct r600_query *r600_context_query_create(struct r600_context *ctx, unsigned
void r600_context_query_destroy(struct r600_context *ctx, struct r600_query *query)
{
pipe_resource_reference((struct pipe_resource**)&query->buffer, NULL);
pipe_resource_reference((struct pipe_resource**)&query->buffer.buf, NULL);
free(query);
}
@@ -1667,27 +1692,30 @@ boolean r600_context_query_result(struct r600_context *ctx,
{
boolean *result_b = (boolean*)vresult;
uint64_t *result_u64 = (uint64_t*)vresult;
union r600_query_result result;
struct pipe_query_data_so_statistics *result_so =
(struct pipe_query_data_so_statistics*)vresult;
if (!r600_query_result(ctx, query, wait))
memset(&result, 0, sizeof(result));
if (!r600_query_result(ctx, query, wait, &result))
return FALSE;
switch (query->type) {
case PIPE_QUERY_OCCLUSION_COUNTER:
case PIPE_QUERY_PRIMITIVES_EMITTED:
case PIPE_QUERY_PRIMITIVES_GENERATED:
*result_u64 = query->result.u64;
*result_u64 = result.u64;
break;
case PIPE_QUERY_OCCLUSION_PREDICATE:
case PIPE_QUERY_SO_OVERFLOW_PREDICATE:
*result_b = query->result.b;
*result_b = result.b;
break;
case PIPE_QUERY_TIME_ELAPSED:
*result_u64 = (1000000 * query->result.u64) / ctx->screen->info.r600_clock_crystal_freq;
*result_u64 = (1000000 * result.u64) / ctx->screen->info.r600_clock_crystal_freq;
break;
case PIPE_QUERY_SO_STATISTICS:
*result_so = query->result.so;
*result_so = result.so;
break;
default:
assert(0);
src/gallium/drivers/r600/r600_query.c
+27 -11
View File
@@ -22,6 +22,7 @@
*/
#include "r600_pipe.h"
#include "r600d.h"
#include "util/u_memory.h"
static struct pipe_query *r600_create_query(struct pipe_context *ctx, unsigned query_type)
{
@@ -56,6 +57,30 @@ static void r600_update_occlusion_query_state(struct r600_context *rctx,
}
}
static void r600_query_discard_results(struct r600_context *rctx,
struct r600_query *query)
{
/* Discard the old query buffers. */
struct r600_query_buffer *prev = query->buffer.previous;
while (prev) {
struct r600_query_buffer *qbuf = prev;
prev = prev->previous;
pipe_resource_reference((struct pipe_resource**)&qbuf->buf, NULL);
FREE(qbuf);
}
/* Obtain a new buffer if the current one can't be mapped without a stall. */
if (rctx->ws->cs_is_buffer_referenced(rctx->cs, query->buffer.buf->cs_buf) ||
rctx->ws->buffer_is_busy(query->buffer.buf->buf, RADEON_USAGE_READWRITE)) {
pipe_resource_reference((struct pipe_resource**)&query->buffer.buf, NULL);
query->buffer.buf = r600_new_query_buffer(rctx, query->type);
}
query->buffer.results_end = 0;
query->buffer.previous = NULL;
}
static void r600_begin_query(struct pipe_context *ctx, struct pipe_query *query)
{
struct r600_context *rctx = (struct r600_context *)ctx;
@@ -63,9 +88,8 @@ static void r600_begin_query(struct pipe_context *ctx, struct pipe_query *query)
r600_update_occlusion_query_state(rctx, rquery->type, 1);
memset(&rquery->result, 0, sizeof(rquery->result));
rquery->results_start = rquery->results_end;
r600_query_begin(rctx, (struct r600_query *)query);
r600_query_discard_results(rctx, rquery);
r600_query_begin(rctx, rquery);
LIST_ADDTAIL(&rquery->list, &rctx->active_query_list);
}
@@ -98,14 +122,6 @@ static void r600_render_condition(struct pipe_context *ctx,
struct r600_query *rquery = (struct r600_query *)query;
int wait_flag = 0;
/* If we already have nonzero result, render unconditionally */
if (query != NULL && rquery->result.u64 != 0) {
if (rctx->current_render_cond) {
r600_render_condition(ctx, NULL, 0);
}
return;
}
rctx->current_render_cond = query;
rctx->current_render_cond_mode = mode;