Commit: 00f203f8fec7401bc5d8f0f359f81df108f423d3
Parent: cb79bfc16f2b08ab723f1968622380f90bef74e7
Author: Randy Palamar
Date: Tue, 3 Mar 2026 10:42:09 -0700
core: avoid crash for unloaded parameter blocks
This sounds dumb but basically the we want to allow the first
block to be lazily allocated and the rest to require allocation.
If we do that we need it may not be loaded when a shader reload
notification comes along despite being labeled as such. As the
blocks are meant to be managed by the compute timeline the
cleanest solution is to just ignore unloaded blocks.
Diffstat:
1 file changed, 3 insertions(+), 9 deletions(-)
diff --git a/beamformer_core.c b/beamformer_core.c
@@ -83,9 +83,6 @@ beamformer_compute_plan_for_block(BeamformerComputeContext *cc, u32 block, Arena
{
assert(block < countof(cc->compute_plans));
BeamformerComputePlan *result = cc->compute_plans[block];
-
- assert(result || arena);
-
if (!result) {
result = SLLPopFreelist(cc->compute_plan_freelist);
if (!result) result = push_struct_no_zero(arena, BeamformerComputePlan);
@@ -1494,12 +1491,9 @@ beamformer_process_input_events(BeamformerCtx *ctx, BeamformerInput *input,
beamformer_reload_shader(ctx, src, ctx->arena, beamformer_shader_names[kind]);
}break;
case BeamformerFileReloadKind_ComputeShader:{
- BeamformerSharedMemory *sm = ctx->shared_memory;
- u32 reserved_blocks = sm->reserved_parameter_blocks;
-
- for (u32 block = 0; block < reserved_blocks; block++) {
- BeamformerComputePlan *cp = beamformer_compute_plan_for_block(&ctx->compute_context, block, 0);
- for (u32 slot = 0; slot < cp->pipeline.shader_count; slot++) {
+ for EachElement(ctx->compute_context.compute_plans, block) {
+ BeamformerComputePlan *cp = ctx->compute_context.compute_plans[block];
+ for (u32 slot = 0; cp && slot < cp->pipeline.shader_count; slot++) {
i32 shader_index = beamformer_shader_reloadable_index_by_shader[cp->pipeline.shaders[slot]];
if (beamformer_reloadable_shader_kinds[shader_index] == frc->compute_shader_kind)
atomic_or_u32(&cp->dirty_programs, 1 << slot);