mirror/rpcs3
1
0
Fork 0
mirror of https://github.com/RPCS3/rpcs3.git synced 2024-11-17 08:11:51 +00:00
Code Issues Packages Projects Releases Wiki Activity

rsx: Improve hit testing when scanning for overlapping surfaces

Browse Source 
- Calculate exact sizes when doing hit tests to avoid false negatives
- Defer page checking until actually require to do memory setup
- Introduce align2 helper to do non-pow2 alignments
This commit is contained in:
kd-11 2019-09-12 17:33:13 +03:00 committed by kd-11
parent 9842823a8c
commit cc313b052f
3 changed files with 46 additions and 58 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

95
rpcs3/Emu/RSX/Common/texture_cache.h
View File

@ -2444,7 +2444,7 @@ namespace rsx
// 1. Invalidate surfaces in range
// 2. Proceed as normal, blit into a 'normal' surface and any upload routines should catch it
m_rtts.invalidate_range(utils::address_range::start_length(dst_address, dst.pitch * dst_h));
use_null_region = (scale_x == 1.f && scale_y == 1.f);
use_null_region = (fcmp(scale_x, 1.f) && fcmp(scale_y, 1.f));
}
// TODO: Handle cases where src or dst can be a depth texture while the other is a color texture - requires a render pass to emulate
@ -2532,6 +2532,10 @@ namespace rsx
areai dst_area = { 0, 0, dst_w, dst_h };
size2i dst_dimensions = { dst.pitch / dst_bpp, dst.height };
const auto src_payload_length = (src.pitch * (src_h - 1) + (src_w * src_bpp));
const auto dst_payload_length = (dst.pitch * (dst_h - 1) + (dst_w * dst_bpp));
if (src_is_render_target)
{
// Attempt to optimize...
@ -2549,23 +2553,6 @@ namespace rsx
{
//LOG_TRACE(RSX, "Blit transfer to surface with dims %dx%d", dst_dimensions.width, dst.height);
}
if (get_location(dst.rsx_address) != CELL_GCM_LOCATION_LOCAL && !dst_is_render_target)
{
// Confirm if the pages actually exist in vm
// Only need to test the extra padding memory
const auto min_end = dst_address + (dst.height * dst.pitch);
const auto max_end = dst.rsx_address + (dst_dimensions.height * dst.pitch);
if (max_end > min_end)
{
if (!vm::check_addr(min_end, (max_end - min_end), vm::page_info_t::page_allocated))
{
// Enforce strict allocation size!
dst_dimensions.height = (dst.clip_height + dst.offset_y);
}
}
}
}
reader_lock lock(m_cache_mutex);
@ -2575,7 +2562,7 @@ namespace rsx
{
// Check for any available region that will fit this one
const auto required_type = (use_null_region) ? texture_upload_context::dma : texture_upload_context::blit_engine_dst;
const auto dst_range = address_range::start_length(dst_address, dst.pitch * dst.clip_height);
const auto dst_range = address_range::start_length(dst_address, dst_payload_length);
auto overlapping_surfaces = find_texture_from_range(dst_range, dst.pitch, required_type);
for (const auto &surface : overlapping_surfaces)
{
@ -2600,6 +2587,8 @@ namespace rsx
cached_dest = surface;
}
// Technically it is totally possible to just extend a pre-existing section
// Will leave this as a TODO
continue;
}
@ -2694,7 +2683,7 @@ namespace rsx
// NOTE: Src address already takes into account the flipped nature of the overlap!
const u32 gcm_format = src_is_argb8 ? CELL_GCM_TEXTURE_A8R8G8B8 : CELL_GCM_TEXTURE_R5G6B5;
const u32 lookup_mask = rsx::texture_upload_context::blit_engine_src | rsx::texture_upload_context::blit_engine_dst | rsx::texture_upload_context::shader_read;
auto overlapping_surfaces = find_texture_from_range<false>(address_range::start_length(src_address, src.pitch * src_h), src.pitch, lookup_mask);
auto overlapping_surfaces = find_texture_from_range<false>(address_range::start_length(src_address, src_payload_length), src.pitch, lookup_mask);
auto old_src_area = src_area;
for (const auto &surface : overlapping_surfaces)
@ -2863,43 +2852,33 @@ namespace rsx
src_area.y2 += scaled_clip_offset_y;
}
// Calculate number of bytes actually modified
u32 mem_base, mem_length;
if (dst_is_render_target)
{
mem_base = dst_address - dst_subres.base_address;
}
else if (cached_dest)
{
mem_base = dst_address - cached_dest->get_section_base();
}
else
{
mem_base = dst_address - dst.rsx_address;
}
if (dst.clip_height == 1)
{
mem_length = dst_w * dst_bpp;
}
else
{
mem_length = (dst.pitch * (dst_h - 1)) + (dst_w * dst_bpp);
}
const auto modified_range = utils::address_range::start_length(dst_address, mem_length);
const auto dst_range = utils::address_range::start_length(dst_address, dst_payload_length);
if (!cached_dest && !dst_is_render_target)
{
verify(HERE), !dest_texture;
// Need to calculate the minium required size that will fit the data, anchored on the rsx_address
// If the application starts off with an 'inseted' section, the guessed dimensions may not fit!
const u32 write_end = dst_address + (dst.pitch * dst_h);
const u32 expected_end = dst.rsx_address + (dst.pitch * dst_dimensions.height);
const u32 write_end = dst_address + dst_payload_length;
u32 block_end = dst.rsx_address + (dst.pitch * dst_dimensions.height);
const u32 section_length = std::max(write_end, expected_end) - dst.rsx_address;
dst_dimensions.height = section_length / dst.pitch;
// Confirm if the pages actually exist in vm
// Only need to test the extra padding memory and only when its on main memory
// NOTE: When src is not a render target, padding is not added speculatively
if (src_is_render_target && get_location(dst.rsx_address) != CELL_GCM_LOCATION_LOCAL)
{
if (block_end > write_end)
{
if (!vm::check_addr(write_end, (block_end - write_end), vm::page_info_t::page_allocated))
{
// Enforce strict allocation size!
block_end = write_end;
}
}
}
const u32 section_length = std::max(write_end, block_end) - dst.rsx_address;
dst_dimensions.height = align2(section_length, dst.pitch) / dst.pitch;
lock.upgrade();
@ -2913,7 +2892,7 @@ namespace rsx
if ((dst_w * dst_bpp) != dst.pitch)
{
// Keep Cell from touching the range we need
const auto prot_range = modified_range.to_page_range();
const auto prot_range = dst_range.to_page_range();
utils::memory_protect(vm::base(prot_range.start), prot_range.length(), utils::protection::no);
force_dma_load = true;
@ -2944,7 +2923,7 @@ namespace rsx
{
// HACK: workaround for data race with Cell
// Pre-lock the memory range we'll be touching, then load with super_ptr
const auto prot_range = modified_range.to_page_range();
const auto prot_range = dst_range.to_page_range();
utils::memory_protect(vm::base(prot_range.start), prot_range.length(), utils::protection::no);
const u16 pitch_in_block = dst.pitch / dst_bpp;
@ -2972,15 +2951,17 @@ namespace rsx
verify(HERE), cached_dest || dst_is_render_target;
// Invalidate any cached subresources in modified range
notify_surface_changed(modified_range);
notify_surface_changed(dst_range);
if (cached_dest)
{
// Validate modified range
u32 mem_offset = dst_address - cached_dest->get_section_base();
verify(HERE), (mem_offset + dst_payload_length) <= cached_dest->get_section_size();
lock.upgrade();
verify(HERE), (mem_base + mem_length) <= cached_dest->get_section_size();
cached_dest->reprotect(utils::protection::no, { mem_base, mem_length });
cached_dest->reprotect(utils::protection::no, { mem_offset, dst_payload_length });
cached_dest->touch(m_cache_update_tag);
update_cache_tag();
}
@ -3045,7 +3026,7 @@ namespace rsx
}
else
{
cached_dest->dma_transfer(cmd, vram_texture, src_area, modified_range, dst.pitch);
cached_dest->dma_transfer(cmd, vram_texture, src_area, dst_range, dst.pitch);
}
blit_op_result result = true;

2
rpcs3/Emu/RSX/VK/VKTexture.cpp
View File

@ -555,7 +555,7 @@ namespace vk
}
else
{
row_pitch = (((layout.width_in_block * block_size_in_bytes) + heap_align - 1) / heap_align) * heap_align;
row_pitch = rsx::align2(layout.width_in_block * block_size_in_bytes, heap_align);
verify(HERE), row_pitch == heap_align;
}

7
rpcs3/Emu/RSX/rsx_utils.h
View File

@ -273,6 +273,13 @@ namespace rsx
CELL_GCM_LOCATION_MAIN;
}
// General purpose alignment without power-of-2 constraint
template <typename T, typename U>
static inline T align2(T value, U alignment)
{
return ((value + alignment - 1) / alignment) * alignment;
}
// Copy memory in inverse direction from source
// Used to scale negatively x axis while transfering image data
template <typename Ts = u8, typename Td = Ts>