The spilling code spills entire vec4 registers regardless of the
components used by the spilled uses.
The inserted stores code force the 4 components, but these loads were
using a variable number of components, causing bugs on loading the
spilled registers.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Qiang Yu <yuq825@gmail.com>
ppir_load_node *load = ppir_node_to_load(load_node);
load->index = -comp->prog->stack_size; /* index sizes are negative */
- load->num_components = src->reg->num_components;
+ load->num_components = 4;
ppir_dest *ld_dest = &load->dest;
ld_dest->type = ppir_target_pipeline;