ilo: allow for different mapping methods

We want to or need to use a different mapping method when when the resource is
busy, the bo format differs from the requested format, and etc.
This commit is contained in:
Chia-I Wu 2013-05-09 14:07:58 +08:00
parent 7cca1aac9d
commit baa44db065
1 changed files with 193 additions and 121 deletions

View File

@ -35,6 +35,23 @@
/* use PIPE_BIND_CUSTOM to indicate MCS */
#define ILO_BIND_MCS PIPE_BIND_CUSTOM
enum ilo_transfer_map_method {
ILO_TRANSFER_MAP_DIRECT,
};
struct ilo_transfer {
struct pipe_transfer base;
enum ilo_transfer_map_method method;
void *ptr;
};
static inline struct ilo_transfer *
ilo_transfer(struct pipe_transfer *transfer)
{
return (struct ilo_transfer *) transfer;
}
static struct intel_bo *
alloc_buf_bo(const struct ilo_resource *res)
{
@ -245,15 +262,138 @@ ilo_transfer_inline_write(struct pipe_context *pipe,
}
static void
ilo_transfer_unmap(struct pipe_context *pipe,
struct pipe_transfer *transfer)
transfer_unmap_direct(struct ilo_context *ilo,
struct ilo_resource *res,
struct ilo_transfer *xfer)
{
struct ilo_resource *res = ilo_resource(transfer->resource);
res->bo->unmap(res->bo);
}
pipe_resource_reference(&transfer->resource, NULL);
FREE(transfer);
static bool
transfer_map_direct(struct ilo_context *ilo,
struct ilo_resource *res,
struct ilo_transfer *xfer)
{
int x, y, err;
if (xfer->base.usage & PIPE_TRANSFER_UNSYNCHRONIZED)
err = res->bo->map_unsynchronized(res->bo);
/* prefer map() when there is the last-level cache */
else if (res->tiling == INTEL_TILING_NONE &&
(ilo->dev->has_llc || (xfer->base.usage & PIPE_TRANSFER_READ)))
err = res->bo->map(res->bo, (xfer->base.usage & PIPE_TRANSFER_WRITE));
else
err = res->bo->map_gtt(res->bo);
if (err)
return false;
/* note that stride is for a block row, not a texel row */
xfer->base.stride = res->bo_stride;
/*
* we can walk through layers when the resource is a texture array or
* when this is the first level of a 3D texture being mapped
*/
if (res->base.array_size > 1 ||
(res->base.target == PIPE_TEXTURE_3D && xfer->base.level == 0)) {
const unsigned qpitch = res->slice_offsets[xfer->base.level][1].y -
res->slice_offsets[xfer->base.level][0].y;
assert(qpitch % res->block_height == 0);
xfer->base.layer_stride = (qpitch / res->block_height) * xfer->base.stride;
}
else {
xfer->base.layer_stride = 0;
}
x = res->slice_offsets[xfer->base.level][xfer->base.box.z].x;
y = res->slice_offsets[xfer->base.level][xfer->base.box.z].y;
x += xfer->base.box.x;
y += xfer->base.box.y;
/* in blocks */
assert(x % res->block_width == 0 && y % res->block_height == 0);
x /= res->block_width;
y /= res->block_height;
xfer->ptr = res->bo->get_virtual(res->bo);
xfer->ptr += y * res->bo_stride + x * res->bo_cpp;
return true;
}
/**
* Choose the best mapping method, depending on the transfer usage and whether
* the bo is busy.
*/
static bool
transfer_map_choose_method(struct ilo_context *ilo,
struct ilo_resource *res,
struct ilo_transfer *xfer)
{
bool will_be_busy, will_stall;
xfer->method = ILO_TRANSFER_MAP_DIRECT;
/* unsynchronized map does not stall */
if (xfer->base.usage & PIPE_TRANSFER_UNSYNCHRONIZED)
return true;
will_be_busy = ilo->cp->bo->references(ilo->cp->bo, res->bo);
if (!will_be_busy) {
/*
* XXX With hardware context support, the bo may be needed by GPU
* without being referenced by ilo->cp->bo. We have to flush
* unconditionally, and that is bad.
*/
if (ilo->cp->hw_ctx)
ilo_cp_flush(ilo->cp);
if (!intel_bo_is_busy(res->bo))
return true;
}
/* bo is busy and mapping it will stall */
will_stall = true;
if (xfer->base.usage & PIPE_TRANSFER_MAP_DIRECTLY) {
/* nothing we can do */
}
else if (xfer->base.usage & PIPE_TRANSFER_DISCARD_WHOLE_RESOURCE) {
/* discard old bo and allocate a new one for mapping */
if (realloc_bo(res))
will_stall = false;
}
else if (xfer->base.usage & PIPE_TRANSFER_FLUSH_EXPLICIT) {
/*
* We could allocate and return a system buffer here. When a region of
* the buffer is explicitly flushed, we pwrite() the region to a
* temporary bo and emit pipelined copy blit.
*
* For now, do nothing.
*/
}
else if (xfer->base.usage & PIPE_TRANSFER_DISCARD_RANGE) {
/*
* We could allocate a temporary bo for mapping, and emit pipelined copy
* blit upon unmapping.
*
* For now, do nothing.
*/
}
if (will_stall) {
if (xfer->base.usage & PIPE_TRANSFER_DONTBLOCK)
return false;
/* flush to make bo busy (so that map() stalls as it should be) */
if (will_be_busy)
ilo_cp_flush(ilo->cp);
}
return true;
}
static void
@ -263,78 +403,25 @@ ilo_transfer_flush_region(struct pipe_context *pipe,
{
}
static bool
map_resource(struct ilo_context *ilo, struct ilo_resource *res,
unsigned usage)
static void
ilo_transfer_unmap(struct pipe_context *pipe,
struct pipe_transfer *transfer)
{
struct ilo_screen *is = ilo_screen(res->base.screen);
bool will_be_busy;
int err;
struct ilo_context *ilo = ilo_context(pipe);
struct ilo_resource *res = ilo_resource(transfer->resource);
struct ilo_transfer *xfer = ilo_transfer(transfer);
/* simply map unsynchronized */
if (usage & PIPE_TRANSFER_UNSYNCHRONIZED) {
err = res->bo->map_unsynchronized(res->bo);
return !err;
switch (xfer->method) {
case ILO_TRANSFER_MAP_DIRECT:
transfer_unmap_direct(ilo, res, xfer);
break;
default:
assert(!"unknown mapping method");
break;
}
/*
* XXX With hardware context support, the bo may be needed by GPU without
* being referenced by ilo->cp->bo. We have to flush unconditionally, and
* that is bad.
*/
if (ilo->cp->hw_ctx)
ilo_cp_flush(ilo->cp);
will_be_busy = ilo->cp->bo->references(ilo->cp->bo, res->bo);
/* see if we can avoid stalling */
if (will_be_busy || intel_bo_is_busy(res->bo)) {
bool will_stall = true;
if (usage & PIPE_TRANSFER_DISCARD_WHOLE_RESOURCE) {
/* discard old bo and allocate a new one for mapping */
if (realloc_bo(res))
will_stall = false;
}
else if (usage & PIPE_TRANSFER_MAP_DIRECTLY) {
/* nothing we can do */
}
else if (usage & PIPE_TRANSFER_FLUSH_EXPLICIT) {
/*
* We could allocate and return a system buffer here. When a region
* of the buffer is explicitly flushed, we pwrite() the region to a
* temporary bo and emit pipelined copy blit.
*
* For now, do nothing.
*/
}
else if (usage & PIPE_TRANSFER_DISCARD_RANGE) {
/*
* We could allocate a temporary bo for mapping, and emit pipelined
* copy blit upon unmapping.
*
* For now, do nothing.
*/
}
if (will_stall) {
if (usage & PIPE_TRANSFER_DONTBLOCK)
return false;
/* flush to make bo busy (so that map() stalls as it should be) */
if (will_be_busy)
ilo_cp_flush(ilo->cp);
}
}
/* prefer map() when there is the last-level cache */
if (res->tiling == INTEL_TILING_NONE &&
(is->dev.has_llc || (usage & PIPE_TRANSFER_READ)))
err = res->bo->map(res->bo, (usage & PIPE_TRANSFER_WRITE));
else
err = res->bo->map_gtt(res->bo);
return !err;
pipe_resource_reference(&xfer->base.resource, NULL);
FREE(xfer);
}
static void *
@ -347,61 +434,46 @@ ilo_transfer_map(struct pipe_context *pipe,
{
struct ilo_context *ilo = ilo_context(pipe);
struct ilo_resource *res = ilo_resource(r);
struct pipe_transfer *xfer;
void *ptr;
int x, y;
struct ilo_transfer *xfer;
int ok;
xfer = MALLOC_STRUCT(pipe_transfer);
if (!xfer)
xfer = MALLOC_STRUCT(ilo_transfer);
if (!xfer) {
*transfer = NULL;
return NULL;
}
if (!map_resource(ilo, res, usage)) {
xfer->base.resource = NULL;
pipe_resource_reference(&xfer->base.resource, &res->base);
xfer->base.level = level;
xfer->base.usage = usage;
xfer->base.box = *box;
ok = transfer_map_choose_method(ilo, res, xfer);
if (ok) {
switch (xfer->method) {
case ILO_TRANSFER_MAP_DIRECT:
ok = transfer_map_direct(ilo, res, xfer);
break;
default:
assert(!"unknown mapping method");
ok = false;
break;
}
}
if (!ok) {
pipe_resource_reference(&xfer->base.resource, NULL);
FREE(xfer);
*transfer = NULL;
return NULL;
}
/* init transfer */
xfer->resource = NULL;
pipe_resource_reference(&xfer->resource, &res->base);
xfer->level = level;
xfer->usage = usage;
xfer->box = *box;
/* stride for a block row, not a texel row */
xfer->stride = res->bo_stride;
*transfer = &xfer->base;
/*
* we can walk through layers when the resource is a texture array or
* when this is the first level of a 3D texture being mapped
*/
if (res->base.array_size > 1 ||
(res->base.target == PIPE_TEXTURE_3D && level == 0)) {
const unsigned qpitch =
res->slice_offsets[level][1].y - res->slice_offsets[level][0].y;
assert(qpitch % res->block_height == 0);
xfer->layer_stride = (qpitch / res->block_height) * xfer->stride;
}
else {
xfer->layer_stride = 0;
}
x = res->slice_offsets[level][box->z].x;
y = res->slice_offsets[level][box->z].y;
x += box->x;
y += box->y;
/* in blocks */
assert(x % res->block_width == 0 && y % res->block_height == 0);
x /= res->block_width;
y /= res->block_height;
ptr = res->bo->get_virtual(res->bo);
ptr += y * res->bo_stride + x * res->bo_cpp;
*transfer = xfer;
return ptr;
return xfer->ptr;
}
static bool