compiler: Save a single copy of the softfp64 shader in the context.

We were recompiling the softfp64 library of functions from GLSL to NIR
every time we compiled a shader that used fp64.  Worse, we were ralloc
stealing it to the GL context.  This meant that we'd accumulate lots of
copies for the lifetime of the context, which was a big space leak.

Instead, we can simply stash a single copy in the GL context, and use
it for subsequent compiles.  Having a single copy should be fine from
a memory context point of view: nir_inline_function_impl already clones
the necessary nir_function_impl's as it inlines.

KHR-GL45.enhanced_layouts.ssb_member_align_non_power_of_2 was previously
OOM'ing a system with 16GB of RAM when using softfp64.  Now it finishes
much more quickly and uses only ~200MB of RAM.

Reviewed-by: Jordan Justen <[email protected]>

1 files changed, 3 insertions, 5 deletions

diff --git a/src/mesa/state_tracker/st_glsl_to_nir.cpp b/src/mesa/state_tracker/st_glsl_to_nir.cpp
index 221c8a030b0..be1fc3b2484 100644
--- a/src/mesa/state_tracker/st_glsl_to_nir.cpp
+++ b/src/mesa/state_tracker/st_glsl_to_nir.cpp
@@ -328,11 +328,9 @@ st_glsl_to_nir(struct st_context *st, struct gl_program *prog,
    }
 
    nir_shader_gather_info(nir, nir_shader_get_entrypoint(nir));
-   nir_shader *softfp64 = NULL;
-   if (nir->info.uses_64bit &&
+   if (!st->ctx->SoftFP64 && nir->info.uses_64bit &&
        (options->lower_doubles_options & nir_lower_fp64_full_software) != 0) {
-      softfp64 = glsl_float64_funcs_to_nir(st->ctx, options);
-      ralloc_steal(ralloc_parent(nir), softfp64);
+      st->ctx->SoftFP64 = glsl_float64_funcs_to_nir(st->ctx, options);
    }
 
    nir_variable_mode mask =
@@ -381,7 +379,7 @@ st_glsl_to_nir(struct st_context *st, struct gl_program *prog,
          }
          if (options->lower_doubles_options) {
             NIR_PASS(progress, nir, nir_lower_doubles,
-                     softfp64, options->lower_doubles_options);
+                     st->ctx->SoftFP64, options->lower_doubles_options);
          }
          NIR_PASS(progress, nir, nir_opt_algebraic);
          lowered_64bit_ops |= progress;