H3cJP/KytyPS5
1
0
Fork 0
mirror of https://github.com/KytyPS5/KytyPS5.git synced 2026-08-03 11:23:49 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

graphics: shader: ignore unused sampler border state

Browse Source 
* Sampler dword 3 only matters when a clamp mode uses border color
  (values >= 4). When no border mode is active, dword 3 is unused
  but can still vary across loop iterations due to wave-lane spills.
  This makes resource tracking think the descriptor is dynamic and
  fail with "unsupported GPU selection".

* Fix by zeroing dword 3 when all clamp modes are non-border.

Signed-off-by: Claxten <claxten10@gmail.com>
This commit is contained in:
Claxten
2026-08-02 02:42:15 +02:00
committed by nmzik
parent 0838142abd
commit 43f30d3ab2
1 changed files with 21 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/graphics/shader/recompiler/ir/ScalarProvenance.cpp
+21 -1
View File
@@ -45,6 +45,8 @@ namespace {
constexpr uint32_t ScalarRegisters = 128;
constexpr uint32_t VectorRegisters = 256;
// Clamp X/Y/Z are consecutive three-bit fields; the high bit of each selects a border mode.
constexpr uint32_t SamplerBorderClampMask = (1u << 2u) | (1u << 5u) | (1u << 8u);
struct ScalarState {
std::array<uint32_t, ScalarRegisters> regs = {};
@@ -647,6 +649,24 @@ private:
return AddDescriptor(descriptor);
}
uint32_t AddSamplerDescriptor(const ScalarState& state, uint32_t base) {
if (base >= ScalarRegisters || 4u > ScalarRegisters - base) {
return ScalarProvenance::Unknown;
}
DescriptorValue descriptor;
descriptor.dword_count = 4;
for (uint32_t i = 0; i < 4; i++) {
descriptor.dwords[i] = state.regs[base + i];
}
if (auto d0 = descriptor.dwords[0];
d0 < m_graph.values.size() && m_graph.values[d0].op == ScalarValueOp::Constant &&
(m_graph.values[d0].imm & SamplerBorderClampMask) == 0) {
// Without a border clamp, the border color and table index in dword 3 are unused.
descriptor.dwords[3] = Constant(0);
}
return AddDescriptor(descriptor);
}
uint32_t AddFlatAddressDescriptor(const Instruction& inst, const ScalarState& state) {
const uint32_t first = FlatStore(inst.op) ? 1u : 0u;
if (inst.src_count < first + 2u) {
@@ -704,7 +724,7 @@ private:
inst.memory.resource_source = AddDescriptor(state, inst.memory.resource * 4u, 8);
if (inst.op == Opcode::ImageSample || inst.op == Opcode::ImageGather4 ||
inst.op == Opcode::ImageGetLod) {
inst.memory.sampler_source = AddDescriptor(state, inst.memory.sampler * 4u, 4);
inst.memory.sampler_source = AddSamplerDescriptor(state, inst.memory.sampler * 4u);
}
}
}