Replace uses of _mesa_bitcount with util_bitcount

and _mesa_bitcount_64 with util_bitcount_64. This fixes a build problem
in nir for platforms that don't have popcount or popcountll, such as
32bit msvc.

v2: - Fix additional uses of _mesa_bitcount added after this was
      originally written

Acked-by: Eric Engestrom <eric.engestrom@intel.com> (v1)
Acked-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>

src/compiler/glsl/ir_constant_expression.cpp

@@ -39,7 +39,7 @@
 #include "ir.h"
 #include "compiler/glsl_types.h"
 #include "util/hash_table.h"
-#include "main/imports.h"
+#include "util/u_math.h"
 
 static float
 dot_f(ir_constant *op0, ir_constant *op1)

src/compiler/glsl/ir_expression_operation.py

@@ -535,7 +535,7 @@ ir_expression_operation = [
 
    # Bit operations, part of ARB_gpu_shader5.
    operation("bitfield_reverse", 1, source_types=(uint_type, int_type), c_expression="bitfield_reverse({src0})"),
-   operation("bit_count", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression="_mesa_bitcount({src0})"),
+   operation("bit_count", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression="util_bitcount({src0})"),
    operation("find_msb", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression={'u': "find_msb_uint({src0})", 'i': "find_msb_int({src0})"}),
    operation("find_lsb", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression="find_msb_uint({src0} & -{src0})"),
 

src/compiler/glsl/link_varyings.cpp

@@ -38,6 +38,7 @@
 #include "link_varyings.h"
 #include "main/macros.h"
 #include "util/hash_table.h"
+#include "util/u_math.h"
 #include "program.h"
 
 
@@ -2879,13 +2880,13 @@ link_varyings(struct gl_shader_program *prog, unsigned first, unsigned last,
 
             /* This must be done after all dead varyings are eliminated. */
             if (sh_i != NULL) {
-               unsigned slots_used = _mesa_bitcount_64(reserved_out_slots);
+               unsigned slots_used = util_bitcount64(reserved_out_slots);
                if (!check_against_output_limit(ctx, prog, sh_i, slots_used)) {
                   return false;
                }
             }
 
-            unsigned slots_used = _mesa_bitcount_64(reserved_in_slots);
+            unsigned slots_used = util_bitcount64(reserved_in_slots);
             if (!check_against_input_limit(ctx, prog, sh_next, slots_used))
                return false;
 

src/compiler/glsl/linker.cpp

@@ -84,6 +84,7 @@
 #include "builtin_functions.h"
 #include "shader_cache.h"
 #include "util/u_string.h"
+#include "util/u_math.h"
 
 #include "main/imports.h"
 #include "main/shaderobj.h"
@@ -3013,8 +3014,8 @@ assign_attribute_or_color_locations(void *mem_ctx,
 
    if (target_index == MESA_SHADER_VERTEX) {
       unsigned total_attribs_size =
-         _mesa_bitcount(used_locations & SAFE_MASK_FROM_INDEX(max_index)) +
+         util_bitcount(used_locations & SAFE_MASK_FROM_INDEX(max_index)) +
-         _mesa_bitcount(double_storage_locations);
+         util_bitcount(double_storage_locations);
       if (total_attribs_size > max_index) {
          linker_error(prog,
                       "attempt to use %d vertex attribute slots only %d available ",
@@ -3077,8 +3078,8 @@ assign_attribute_or_color_locations(void *mem_ctx,
     */
    if (target_index == MESA_SHADER_VERTEX) {
       unsigned total_attribs_size =
-         _mesa_bitcount(used_locations & SAFE_MASK_FROM_INDEX(max_index)) +
+         util_bitcount(used_locations & SAFE_MASK_FROM_INDEX(max_index)) +
-         _mesa_bitcount(double_storage_locations);
+         util_bitcount(double_storage_locations);
       if (total_attribs_size > max_index) {
          linker_error(prog,
                       "attempt to use %d vertex attribute slots only %d available ",

src/compiler/nir/nir.c

@@ -31,8 +31,8 @@
 #include <limits.h>
 #include <assert.h>
 #include <math.h>
+#include "util/u_math.h"
 
-#include "main/imports.h" /* _mesa_bitcount_64 */
 #include "main/menums.h" /* BITFIELD64_MASK */
 
 nir_shader *
@@ -1862,7 +1862,7 @@ nir_system_value_from_intrinsic(nir_intrinsic_op intrin)
  * the original OpenGL single-slot input numbering.  The mapping from old
  * locations to new locations is as follows:
  *
- *    new_loc = loc + _mesa_bitcount(dual_slot & BITFIELD64_MASK(loc))
+ *    new_loc = loc + util_bitcount(dual_slot & BITFIELD64_MASK(loc))
  */
 void
 nir_remap_dual_slot_attributes(nir_shader *shader, uint64_t *dual_slot)
@@ -1879,7 +1879,7 @@ nir_remap_dual_slot_attributes(nir_shader *shader, uint64_t *dual_slot)
 
    nir_foreach_variable(var, &shader->inputs) {
       var->data.location +=
-         _mesa_bitcount_64(*dual_slot & BITFIELD64_MASK(var->data.location));
+         util_bitcount64(*dual_slot & BITFIELD64_MASK(var->data.location));
    }
 }
 

src/compiler/nir/nir_split_vars.c

@@ -26,8 +26,8 @@
 #include "nir_deref.h"
 #include "nir_vla.h"
 
-/* Needed for _mesa_bitcount() */
+#include "util/u_math.h"
-#include "main/macros.h"
+
 
 struct split_var_state {
    void *mem_ctx;
@@ -1277,7 +1277,7 @@ shrink_vec_var_list(struct exec_list *vars,
       }
 
       /* Build the new var type */
-      unsigned new_num_comps = _mesa_bitcount(usage->comps_kept);
+      unsigned new_num_comps = util_bitcount(usage->comps_kept);
       const struct glsl_type *new_type =
          glsl_vector_type(glsl_get_base_type(vec_type), new_num_comps);
       for (int i = usage->num_levels - 1; i >= 0; i--) {

src/gallium/state_trackers/glx/xlib/glx_api.c

@@ -41,6 +41,7 @@
 #include "xm_api.h"
 #include "main/imports.h"
 #include "main/errors.h"
+#include "util/u_math.h"
 
 /* An "Atrribs/Attribs" typo was fixed in glxproto.h in Nov 2014.
  * This is in case we don't have the updated header.
@@ -419,9 +420,9 @@ get_visual( Display *dpy, int scr, unsigned int depth, int xclass )
     * 10 bits per color channel.  Mesa's limited to a max of 8 bits/channel.
     */
    if (vis && depth > 24 && (xclass==TrueColor || xclass==DirectColor)) {
-      if (_mesa_bitcount((GLuint) vis->red_mask  ) <= 8 &&
+      if (util_bitcount((GLuint) vis->red_mask  ) <= 8 &&
-          _mesa_bitcount((GLuint) vis->green_mask) <= 8 &&
+          util_bitcount((GLuint) vis->green_mask) <= 8 &&
-          _mesa_bitcount((GLuint) vis->blue_mask ) <= 8) {
+          util_bitcount((GLuint) vis->blue_mask ) <= 8) {
          return vis;
       }
       else {

src/gallium/state_trackers/glx/xlib/xm_api.c

@@ -65,6 +65,7 @@
 
 #include "util/u_atomic.h"
 #include "util/u_inlines.h"
+#include "util/u_math.h"
 
 #include "hud/hud_context.h"
 
@@ -825,9 +826,9 @@ XMesaVisual XMesaCreateVisual( Display *display,
    {
       const int xclass = v->visualType;
       if (xclass == GLX_TRUE_COLOR || xclass == GLX_DIRECT_COLOR) {
-         red_bits   = _mesa_bitcount(GET_REDMASK(v));
+         red_bits   = util_bitcount(GET_REDMASK(v));
-         green_bits = _mesa_bitcount(GET_GREENMASK(v));
+         green_bits = util_bitcount(GET_GREENMASK(v));
-         blue_bits  = _mesa_bitcount(GET_BLUEMASK(v));
+         blue_bits  = util_bitcount(GET_BLUEMASK(v));
       }
       else {
          /* this is an approximation */
@@ -1180,8 +1181,8 @@ XMesaCreatePixmapTextureBuffer(XMesaVisual v, Pixmap p,
       if (ctx->Extensions.ARB_texture_non_power_of_two) {
          target = GLX_TEXTURE_2D_EXT;
       }
-      else if (   _mesa_bitcount(b->width)  == 1
+      else if (   util_bitcount(b->width)  == 1
-               && _mesa_bitcount(b->height) == 1) {
+               && util_bitcount(b->height) == 1) {
          /* power of two size */
          if (b->height == 1) {
             target = GLX_TEXTURE_1D_EXT;

src/gallium/targets/libgl-xlib/Makefile.am

@@ -62,6 +62,7 @@ lib@GL_LIB@_la_LIBADD = \
 	$(top_builddir)/src/mapi/glapi/libglapi.la \
 	$(top_builddir)/src/mesa/libmesagallium.la \
 	$(top_builddir)/src/gallium/auxiliary/libgallium.la \
+	$(top_builddir)/src/util/libmesautil.la \
 	$(SHARED_GLAPI_LIB) \
 	$(GL_LIB_DEPS) \
 	$(CLOCK_LIB) \

src/intel/blorp/blorp_blit.c

@@ -29,6 +29,7 @@
 #include "util/format_rgb9e5.h"
 /* header-only include needed for _mesa_unorm_to_float and friends. */
 #include "mesa/main/format_utils.h"
+#include "util/u_math.h"
 
 #define FILE_DEBUG_FLAG DEBUG_BLORP
 
@@ -582,7 +583,7 @@ static inline int count_trailing_one_bits(unsigned value)
 #ifdef HAVE___BUILTIN_CTZ
    return __builtin_ctz(~value);
 #else
-   return _mesa_bitcount(value & ~(value + 1));
+   return util_bitcount(value & ~(value + 1));
 #endif
 }
 
@@ -634,7 +635,7 @@ blorp_nir_manual_blend_average(nir_builder *b, struct brw_blorp_blit_vars *v,
    nir_ssa_def *texture_data[5];
    unsigned stack_depth = 0;
    for (unsigned i = 0; i < tex_samples; ++i) {
-      assert(stack_depth == _mesa_bitcount(i)); /* Loop invariant */
+      assert(stack_depth == util_bitcount(i)); /* Loop invariant */
 
       /* Push sample i onto the stack */
       assert(stack_depth < ARRAY_SIZE(texture_data));

src/intel/compiler/brw_fs.cpp

@@ -39,6 +39,7 @@
 #include "compiler/glsl_types.h"
 #include "compiler/nir/nir_builder.h"
 #include "program/prog_parameter.h"
+#include "util/u_math.h"
 
 using namespace brw;
 
@@ -1534,7 +1535,7 @@ fs_visitor::calculate_urb_setup()
    int urb_next = 0;
    /* Figure out where each of the incoming setup attributes lands. */
    if (devinfo->gen >= 6) {
-      if (_mesa_bitcount_64(nir->info.inputs_read &
+      if (util_bitcount64(nir->info.inputs_read &
                             BRW_FS_VARYING_INPUT_MASK) <= 16) {
          /* The SF/SBE pipeline stage can do arbitrary rearrangement of the
           * first 16 varying inputs, so we can put them wherever we want.

src/intel/compiler/brw_fs_nir.cpp

@@ -25,6 +25,7 @@
 #include "brw_fs.h"
 #include "brw_fs_surface_builder.h"
 #include "brw_nir.h"
+#include "util/u_math.h"
 
 using namespace brw;
 using namespace brw::surface_access;
@@ -751,7 +752,7 @@ fs_visitor::nir_emit_alu(const fs_builder &bld, nir_alu_instr *instr)
       /* Since NIR is doing the scalarizing for us, we should only ever see
        * vectorized operations with a single channel.
        */
-      assert(_mesa_bitcount(instr->dest.write_mask) == 1);
+      assert(util_bitcount(instr->dest.write_mask) == 1);
       channel = ffs(instr->dest.write_mask) - 1;
 
       result = offset(result, bld, channel);

src/intel/compiler/brw_nir.c

@@ -26,6 +26,7 @@
 #include "common/gen_debug.h"
 #include "compiler/glsl_types.h"
 #include "compiler/nir/nir_builder.h"
+#include "util/u_math.h"
 
 static bool
 is_input(nir_intrinsic_instr *intrin)
@@ -243,7 +244,7 @@ brw_nir_lower_vs_inputs(nir_shader *nir,
        BITFIELD64_BIT(SYSTEM_VALUE_VERTEX_ID_ZERO_BASE) |
        BITFIELD64_BIT(SYSTEM_VALUE_INSTANCE_ID));
 
-   const unsigned num_inputs = _mesa_bitcount_64(nir->info.inputs_read);
+   const unsigned num_inputs = util_bitcount64(nir->info.inputs_read);
 
    nir_foreach_function(function, nir) {
       if (!function->impl)
@@ -322,7 +323,7 @@ brw_nir_lower_vs_inputs(nir_shader *nir,
                 * before it and counting the bits.
                 */
                int attr = nir_intrinsic_base(intrin);
-               int slot = _mesa_bitcount_64(nir->info.inputs_read &
+               int slot = util_bitcount64(nir->info.inputs_read &
                                             BITFIELD64_MASK(attr));
                nir_intrinsic_set_base(intrin, slot);
                break;

src/intel/compiler/brw_vec4.cpp

@@ -31,6 +31,7 @@
 #include "brw_dead_control_flow.h"
 #include "common/gen_debug.h"
 #include "program/prog_parameter.h"
+#include "util/u_math.h"
 
 #define MAX_INSTRUCTION (1 << 30)
 
@@ -2845,7 +2846,7 @@ brw_compile_vs(const struct brw_compiler *compiler, void *log_data,
       ((1 << shader->info.cull_distance_array_size) - 1) <<
       shader->info.clip_distance_array_size;
 
-   unsigned nr_attribute_slots = _mesa_bitcount_64(prog_data->inputs_read);
+   unsigned nr_attribute_slots = util_bitcount64(prog_data->inputs_read);
 
    /* gl_VertexID and gl_InstanceID are system values, but arrive via an
     * incoming vertex attribute.  So, add an extra slot.

src/intel/compiler/brw_vec4_visitor.cpp

@@ -24,6 +24,7 @@
 #include "brw_vec4.h"
 #include "brw_cfg.h"
 #include "brw_eu.h"
+#include "util/u_math.h"
 
 namespace brw {
 
@@ -1317,7 +1318,7 @@ vec4_visitor::emit_urb_slot(dst_reg reg, int varying)
        * determine which edges should be drawn as wireframe.
        */
       current_annotation = "edge flag";
-      int edge_attr = _mesa_bitcount_64(nir->info.inputs_read &
+      int edge_attr = util_bitcount64(nir->info.inputs_read &
                                         BITFIELD64_MASK(VERT_ATTRIB_EDGEFLAG));
       emit(MOV(reg, src_reg(dst_reg(ATTR, edge_attr,
                                     glsl_type::float_type, WRITEMASK_XYZW))));

src/intel/vulkan/anv_blorp.c

@@ -296,7 +296,7 @@ void anv_CmdCopyImage(
 
       assert(anv_image_aspects_compatible(src_mask, dst_mask));
 
-      if (_mesa_bitcount(src_mask) > 1) {
+      if (util_bitcount(src_mask) > 1) {
          uint32_t aspect_bit;
          anv_foreach_image_aspect_bit(aspect_bit, src_image, src_mask) {
             struct blorp_surf src_surf, dst_surf;

src/intel/vulkan/anv_image.c

@@ -32,6 +32,7 @@
 #include "anv_private.h"
 #include "util/debug.h"
 #include "vk_util.h"
+#include "util/u_math.h"
 
 #include "vk_format_info.h"
 
@@ -814,7 +815,7 @@ anv_layout_to_aux_usage(const struct gen_device_info * const devinfo,
    assert(image != NULL);
 
    /* The aspect must be exactly one of the image aspects. */
-   assert(_mesa_bitcount(aspect) == 1 && (aspect & image->aspects));
+   assert(util_bitcount(aspect) == 1 && (aspect & image->aspects));
 
    /* Determine the optimal buffer. */
 
@@ -942,7 +943,7 @@ anv_layout_to_fast_clear_type(const struct gen_device_info * const devinfo,
                               const VkImageLayout layout)
 {
    /* The aspect must be exactly one of the image aspects. */
-   assert(_mesa_bitcount(aspect) == 1 && (aspect & image->aspects));
+   assert(util_bitcount(aspect) == 1 && (aspect & image->aspects));
 
    uint32_t plane = anv_image_aspect_to_plane(image->aspects, aspect);
 
@@ -1230,11 +1231,11 @@ static VkImageAspectFlags
 remap_aspect_flags(VkImageAspectFlags view_aspects)
 {
    if (view_aspects & VK_IMAGE_ASPECT_ANY_COLOR_BIT_ANV) {
-      if (_mesa_bitcount(view_aspects) == 1)
+      if (util_bitcount(view_aspects) == 1)
          return VK_IMAGE_ASPECT_COLOR_BIT;
 
       VkImageAspectFlags color_aspects = 0;
-      for (uint32_t i = 0; i < _mesa_bitcount(view_aspects); i++)
+      for (uint32_t i = 0; i < util_bitcount(view_aspects); i++)
          color_aspects |= VK_IMAGE_ASPECT_PLANE_0_BIT << i;
       return color_aspects;
    }

src/intel/vulkan/anv_nir_lower_multiview.c

@@ -57,7 +57,7 @@ build_instance_id(struct lower_multiview_state *state)
        */
       state->instance_id =
          nir_idiv(b, nir_load_instance_id(b),
-                     nir_imm_int(b, _mesa_bitcount(state->view_mask)));
+                     nir_imm_int(b, util_bitcount(state->view_mask)));
    }
 
    return state->instance_id;
@@ -72,7 +72,7 @@ build_view_index(struct lower_multiview_state *state)
       b->cursor = nir_before_block(nir_start_block(b->impl));
 
       assert(state->view_mask != 0);
-      if (_mesa_bitcount(state->view_mask) == 1) {
+      if (util_bitcount(state->view_mask) == 1) {
          /* Set the view index directly. */
          state->view_index = nir_imm_int(b, ffs(state->view_mask) - 1);
       } else if (state->builder.shader->info.stage == MESA_SHADER_VERTEX) {
@@ -85,7 +85,7 @@ build_view_index(struct lower_multiview_state *state)
           */
          nir_ssa_def *compacted =
             nir_umod(b, nir_load_instance_id(b),
-                        nir_imm_int(b, _mesa_bitcount(state->view_mask)));
+                        nir_imm_int(b, util_bitcount(state->view_mask)));
 
          if (util_is_power_of_two_or_zero(state->view_mask + 1)) {
             /* If we have a full view mask, then compacted is what we want */
@@ -206,7 +206,7 @@ anv_nir_lower_multiview(nir_shader *shader, uint32_t view_mask)
 
       /* Unless there is only one possible view index (that would be set
        * directly), pass it to the next stage. */
-      if (_mesa_bitcount(state.view_mask) != 1) {
+      if (util_bitcount(state.view_mask) != 1) {
          nir_variable *view_index_out =
             nir_variable_create(shader, nir_var_shader_out,
                                 glsl_int_type(), "view index");

src/intel/vulkan/anv_pipeline.c

@@ -364,7 +364,7 @@ populate_wm_prog_key(const struct gen_device_info *devinfo,
          key->color_outputs_valid |= (1 << i);
    }
 
-   key->nr_color_regions = _mesa_bitcount(key->color_outputs_valid);
+   key->nr_color_regions = util_bitcount(key->color_outputs_valid);
 
    key->replicate_alpha = key->nr_color_regions > 1 &&
                           ms_info && ms_info->alphaToCoverageEnable;

src/intel/vulkan/anv_private.h

@@ -54,6 +54,7 @@
 #include "util/set.h"
 #include "util/u_atomic.h"
 #include "util/u_vector.h"
+#include "util/u_math.h"
 #include "util/vma.h"
 #include "vk_alloc.h"
 #include "vk_debug_report.h"
@@ -2557,7 +2558,7 @@ anv_plane_to_aspect(VkImageAspectFlags image_aspects,
                     uint32_t plane)
 {
    if (image_aspects & VK_IMAGE_ASPECT_ANY_COLOR_BIT_ANV) {
-      if (_mesa_bitcount(image_aspects) > 1)
+      if (util_bitcount(image_aspects) > 1)
          return VK_IMAGE_ASPECT_PLANE_0_BIT << plane;
       return VK_IMAGE_ASPECT_COLOR_BIT;
    }
@@ -2968,7 +2969,7 @@ anv_image_aspects_compatible(VkImageAspectFlags aspects1,
    /* Only 1 color aspects are compatibles. */
    if ((aspects1 & VK_IMAGE_ASPECT_ANY_COLOR_BIT_ANV) != 0 &&
        (aspects2 & VK_IMAGE_ASPECT_ANY_COLOR_BIT_ANV) != 0 &&
-       _mesa_bitcount(aspects1) == _mesa_bitcount(aspects2))
+       util_bitcount(aspects1) == util_bitcount(aspects2))
       return true;
 
    return false;
@@ -3179,7 +3180,7 @@ struct anv_subpass {
 static inline unsigned
 anv_subpass_view_count(const struct anv_subpass *subpass)
 {
-   return MAX2(1, _mesa_bitcount(subpass->view_mask));
+   return MAX2(1, util_bitcount(subpass->view_mask));
 }
 
 struct anv_render_pass_attachment {

src/intel/vulkan/genX_cmd_buffer.c

@@ -787,7 +787,7 @@ genX(cmd_buffer_mark_image_written)(struct anv_cmd_buffer *cmd_buffer,
                                     uint32_t layer_count)
 {
    /* The aspect must be exactly one of the image aspects. */
-   assert(_mesa_bitcount(aspect) == 1 && (aspect & image->aspects));
+   assert(util_bitcount(aspect) == 1 && (aspect & image->aspects));
 
    /* The only compression types with more than just fast-clears are MCS,
     * CCS_E, and HiZ.  With HiZ we just trust the layout and don't actually
@@ -1859,7 +1859,7 @@ cmd_buffer_alloc_push_constants(struct anv_cmd_buffer *cmd_buffer)
 #endif
 
    const unsigned num_stages =
-      _mesa_bitcount(stages & VK_SHADER_STAGE_ALL_GRAPHICS);
+      util_bitcount(stages & VK_SHADER_STAGE_ALL_GRAPHICS);
    unsigned size_per_stage = push_constant_kb / num_stages;
 
    /* Broadwell+ and Haswell gt3 require that the push constant sizes be in

src/intel/vulkan/genX_query.c

@@ -70,7 +70,7 @@ VkResult genX(CreateQueryPool)(
       pipeline_statistics &= ANV_PIPELINE_STATISTICS_MASK;
 
       /* Statistics queries have a min and max for every statistic */
-      uint64s_per_slot += 2 * _mesa_bitcount(pipeline_statistics);
+      uint64s_per_slot += 2 * util_bitcount(pipeline_statistics);
       break;
    default:
       assert(!"Invalid query type");
@@ -272,7 +272,7 @@ VkResult genX(GetQueryPoolResults)(
 
                idx++;
             }
-            assert(idx == _mesa_bitcount(pool->pipeline_statistics));
+            assert(idx == util_bitcount(pool->pipeline_statistics));
             break;
          }
 
@@ -289,7 +289,7 @@ VkResult genX(GetQueryPoolResults)(
 
       if (flags & VK_QUERY_RESULT_WITH_AVAILABILITY_BIT) {
          uint32_t idx = (pool->type == VK_QUERY_TYPE_PIPELINE_STATISTICS) ?
-                        _mesa_bitcount(pool->pipeline_statistics) : 1;
+                        util_bitcount(pool->pipeline_statistics) : 1;
          cpu_write_query_result(pData, flags, idx, available);
       }
 
@@ -489,7 +489,7 @@ void genX(CmdEndQuery)(
     */
    if (cmd_buffer->state.subpass && cmd_buffer->state.subpass->view_mask) {
       const uint32_t num_queries =
-         _mesa_bitcount(cmd_buffer->state.subpass->view_mask);
+         util_bitcount(cmd_buffer->state.subpass->view_mask);
       if (num_queries > 1)
          emit_zero_queries(cmd_buffer, pool, query + 1, num_queries - 1);
    }
@@ -546,7 +546,7 @@ void genX(CmdWriteTimestamp)(
     */
    if (cmd_buffer->state.subpass && cmd_buffer->state.subpass->view_mask) {
       const uint32_t num_queries =
-         _mesa_bitcount(cmd_buffer->state.subpass->view_mask);
+         util_bitcount(cmd_buffer->state.subpass->view_mask);
       if (num_queries > 1)
          emit_zero_queries(cmd_buffer, pool, query + 1, num_queries - 1);
    }
@@ -778,7 +778,7 @@ void genX(CmdCopyQueryPoolResults)(
 
             idx++;
          }
-         assert(idx == _mesa_bitcount(pool->pipeline_statistics));
+         assert(idx == util_bitcount(pool->pipeline_statistics));
          break;
       }
 
@@ -795,7 +795,7 @@ void genX(CmdCopyQueryPoolResults)(
 
       if (flags & VK_QUERY_RESULT_WITH_AVAILABILITY_BIT) {
          uint32_t idx = (pool->type == VK_QUERY_TYPE_PIPELINE_STATISTICS) ?
-                        _mesa_bitcount(pool->pipeline_statistics) : 1;
+                        util_bitcount(pool->pipeline_statistics) : 1;
 
          emit_load_alu_reg_u64(&cmd_buffer->batch, CS_GPR(0),
                                &pool->bo, slot_offset);

src/mesa/drivers/common/meta.c

@@ -88,6 +88,7 @@
 #include "util/bitscan.h"
 #include "util/ralloc.h"
 #include "compiler/nir/nir.h"
+#include "util/u_math.h"
 
 /** Return offset in bytes of the field within a vertex struct */
 #define OFFSET(FIELD) ((void *) offsetof(struct vertex, FIELD))
@@ -1611,7 +1612,7 @@ _mesa_meta_drawbuffers_from_bitfield(GLbitfield bits)
    assert((bits & ~BUFFER_BITS_COLOR) == 0);
 
    /* Make sure we don't overflow any arrays. */
-   assert(_mesa_bitcount(bits) <= MAX_DRAW_BUFFERS);
+   assert(util_bitcount(bits) <= MAX_DRAW_BUFFERS);
 
    enums[0] = GL_NONE;
 
@@ -1651,7 +1652,7 @@ _mesa_meta_drawbuffers_and_colormask(struct gl_context *ctx, GLbitfield mask)
    assert((mask & ~BUFFER_BITS_COLOR) == 0);
 
    /* Make sure we don't overflow any arrays. */
-   assert(_mesa_bitcount(mask) <= MAX_DRAW_BUFFERS);
+   assert(util_bitcount(mask) <= MAX_DRAW_BUFFERS);
 
    enums[0] = GL_NONE;
 

src/mesa/drivers/dri/i965/brw_curbe.c

@@ -64,6 +64,7 @@
 #include "brw_defines.h"
 #include "brw_state.h"
 #include "brw_util.h"
+#include "util/u_math.h"
 
 
 /**
@@ -86,7 +87,7 @@ static void calculate_curbe_offsets( struct brw_context *brw )
 
    /* _NEW_TRANSFORM */
    if (ctx->Transform.ClipPlanesEnabled) {
-      GLuint nr_planes = 6 + _mesa_bitcount(ctx->Transform.ClipPlanesEnabled);
+      GLuint nr_planes = 6 + util_bitcount(ctx->Transform.ClipPlanesEnabled);
       nr_clip_regs = (nr_planes * 4 + 15) / 16;
    }
 

src/mesa/drivers/dri/i965/brw_performance_query.c

@@ -67,6 +67,7 @@
 #include "util/ralloc.h"
 #include "util/hash_table.h"
 #include "util/list.h"
+#include "util/u_math.h"
 
 #include "brw_context.h"
 #include "brw_defines.h"
@@ -1985,11 +1986,11 @@ compute_topology_builtins(struct brw_context *brw)
 
    for (int i = 0; i < sizeof(devinfo->subslice_masks[i]); i++) {
       brw->perfquery.sys_vars.n_eu_sub_slices +=
-         _mesa_bitcount(devinfo->subslice_masks[i]);
+         util_bitcount(devinfo->subslice_masks[i]);
    }
 
    for (int i = 0; i < sizeof(devinfo->eu_masks); i++)
-      brw->perfquery.sys_vars.n_eus += _mesa_bitcount(devinfo->eu_masks[i]);
+      brw->perfquery.sys_vars.n_eus += util_bitcount(devinfo->eu_masks[i]);
 
    brw->perfquery.sys_vars.eu_threads_count =
       brw->perfquery.sys_vars.n_eus * devinfo->num_thread_per_eu;

src/mesa/drivers/dri/i965/brw_wm.c

@@ -40,6 +40,7 @@
 #include "brw_program.h"
 
 #include "util/ralloc.h"
+#include "util/u_math.h"
 
 static void
 assign_fs_binding_table_offsets(const struct gen_device_info *devinfo,
@@ -559,7 +560,7 @@ brw_wm_populate_key(struct brw_context *brw, struct brw_wm_prog_key *key)
    }
 
    /* BRW_NEW_VUE_MAP_GEOM_OUT */
-   if (devinfo->gen < 6 || _mesa_bitcount_64(prog->info.inputs_read &
+   if (devinfo->gen < 6 || util_bitcount64(prog->info.inputs_read &
                                              BRW_FS_VARYING_INPUT_MASK) > 16) {
       key->input_slots_valid = brw->vue_map_geom_out.slots_valid;
    }
@@ -632,14 +633,14 @@ brw_wm_populate_default_key(const struct gen_device_info *devinfo,
       key->iz_lookup |= BRW_WM_IZ_DEPTH_WRITE_ENABLE_BIT;
    }
 
-   if (devinfo->gen < 6 || _mesa_bitcount_64(prog->info.inputs_read &
+   if (devinfo->gen < 6 || util_bitcount64(prog->info.inputs_read &
                                              BRW_FS_VARYING_INPUT_MASK) > 16) {
       key->input_slots_valid = prog->info.inputs_read | VARYING_BIT_POS;
    }
 
    brw_setup_tex_for_precompile(devinfo, &key->tex, prog);
 
-   key->nr_color_regions = _mesa_bitcount_64(outputs_written &
+   key->nr_color_regions = util_bitcount64(outputs_written &
          ~(BITFIELD64_BIT(FRAG_RESULT_DEPTH) |
            BITFIELD64_BIT(FRAG_RESULT_STENCIL) |
            BITFIELD64_BIT(FRAG_RESULT_SAMPLE_MASK)));

src/mesa/drivers/x11/Makefile.am

@@ -68,6 +68,7 @@ GL_PATCH = 0
 lib@GL_LIB@_la_LIBADD = \
 	$(top_builddir)/src/mesa/libmesa.la \
 	$(top_builddir)/src/mapi/glapi/libglapi.la \
+	$(top_builddir)/src/util/libmesautil.la \
 	$(SHARED_GLAPI_LIB) \
 	$(GL_LIB_DEPS)
 

src/mesa/drivers/x11/fakeglx.c

@@ -52,6 +52,7 @@
 #include "main/version.h"
 #include "xfonts.h"
 #include "xmesaP.h"
+#include "util/u_math.h"
 
 /* This indicates the client-side GLX API and GLX encoder version. */
 #define CLIENT_MAJOR_VERSION 1
@@ -510,9 +511,9 @@ get_visual( Display *dpy, int scr, unsigned int depth, int xclass )
     * 10 bits per color channel.  Mesa's limited to a max of 8 bits/channel.
     */
    if (vis && depth > 24 && (xclass==TrueColor || xclass==DirectColor)) {
-      if (_mesa_bitcount((GLuint) vis->red_mask  ) <= 8 &&
+      if (util_bitcount((GLuint) vis->red_mask  ) <= 8 &&
-          _mesa_bitcount((GLuint) vis->green_mask) <= 8 &&
+          util_bitcount((GLuint) vis->green_mask) <= 8 &&
-          _mesa_bitcount((GLuint) vis->blue_mask ) <= 8) {
+          util_bitcount((GLuint) vis->blue_mask ) <= 8) {
          return vis;
       }
       else {

src/mesa/drivers/x11/meson.build

@@ -32,7 +32,7 @@ libgl = shared_library(
   include_directories : [
     inc_include, inc_src, inc_mesa, inc_mapi, inc_gallium, inc_gallium_aux
   ],
-  link_with : [libmesa_classic, libglapi_static, gl_link_with],
+  link_with : [libmesa_classic, libglapi_static, libmesa_util, gl_link_with],
   dependencies : [dep_x11, dep_xext, dep_xcb, dep_thread],
   version : '1.6.0',
   install : true,

src/mesa/drivers/x11/xm_api.c

@@ -84,6 +84,7 @@
 #include "tnl/t_pipeline.h"
 #include "drivers/common/driverfuncs.h"
 #include "drivers/common/meta.h"
+#include "util/u_math.h"
 
 /**
  * Global X driver lock
@@ -463,9 +464,9 @@ setup_truecolor(XMesaVisual v, XMesaBuffer buffer, XMesaColormap cmap)
           3*16, 11*16,  1*16,  9*16,
          15*16,  7*16, 13*16,  5*16,
       };
-      GLint rBits = _mesa_bitcount(rmask);
+      GLint rBits = util_bitcount(rmask);
-      GLint gBits = _mesa_bitcount(gmask);
+      GLint gBits = util_bitcount(gmask);
-      GLint bBits = _mesa_bitcount(bmask);
+      GLint bBits = util_bitcount(bmask);
       GLint maxBits;
       GLuint i;
 
@@ -828,9 +829,9 @@ XMesaVisual XMesaCreateVisual( XMesaDisplay *display,
    {
       const int xclass = v->visualType;
       if (xclass == GLX_TRUE_COLOR || xclass == GLX_DIRECT_COLOR) {
-         red_bits   = _mesa_bitcount(GET_REDMASK(v));
+         red_bits   = util_bitcount(GET_REDMASK(v));
-         green_bits = _mesa_bitcount(GET_GREENMASK(v));
+         green_bits = util_bitcount(GET_GREENMASK(v));
-         blue_bits  = _mesa_bitcount(GET_BLUEMASK(v));
+         blue_bits  = util_bitcount(GET_BLUEMASK(v));
       }
       else {
          /* this is an approximation */
@@ -1095,8 +1096,8 @@ XMesaCreatePixmapTextureBuffer(XMesaVisual v, XMesaPixmap p,
       if (ctx->Extensions.ARB_texture_non_power_of_two) {
          target = GLX_TEXTURE_2D_EXT;
       }
-      else if (   _mesa_bitcount(width)  == 1
+      else if (   util_bitcount(width)  == 1
-               && _mesa_bitcount(height) == 1) {
+               && util_bitcount(height) == 1) {
          /* power of two size */
          if (height == 1) {
             target = GLX_TEXTURE_1D_EXT;

src/mesa/main/arrayobj.c

@@ -53,6 +53,7 @@
 #include "varray.h"
 #include "util/bitscan.h"
 #include "util/u_atomic.h"
+#include "util/u_math.h"
 
 
 const GLubyte
@@ -755,7 +756,7 @@ _mesa_update_vao_derived_arrays(struct gl_context *ctx,
           * grouping information in a seperate array beside
           * gl_array_attributes/gl_vertex_buffer_binding.
           */
-         assert(_mesa_bitcount(binding->_BoundArrays & vao->_Enabled) == 1
+         assert(util_bitcount(binding->_BoundArrays & vao->_Enabled) == 1
                 || (vao->_Enabled & ~binding->_BoundArrays) == 0);
 
          /* Start this current effective binding with the array */
@@ -775,7 +776,7 @@ _mesa_update_vao_derived_arrays(struct gl_context *ctx,
                &vao->BufferBinding[attrib2->BufferBindingIndex];
 
             /* See the comment at the same assert above. */
-            assert(_mesa_bitcount(binding2->_BoundArrays & vao->_Enabled) == 1
+            assert(util_bitcount(binding2->_BoundArrays & vao->_Enabled) == 1
                    || (vao->_Enabled & ~binding->_BoundArrays) == 0);
 
             /* Check if we have an identical binding */

src/mesa/main/buffers.c

@@ -37,6 +37,7 @@
 #include "fbobject.h"
 #include "mtypes.h"
 #include "util/bitscan.h"
+#include "util/u_math.h"
 
 
 #define BAD_MASK ~0u
@@ -468,7 +469,7 @@ draw_buffers(struct gl_context *ctx, struct gl_framebuffer *fb, GLsizei n,
           * For OpenGL 4.x we check that behaviour. For any previous version we
           * keep considering it wrong (as INVALID_ENUM).
           */
-         if (_mesa_bitcount(destMask[output]) > 1) {
+         if (util_bitcount(destMask[output]) > 1) {
             if (_mesa_is_winsys_fbo(fb) && ctx->Version >= 40 &&
                 buffers[output] == GL_BACK) {
                if (n != 1) {
@@ -722,7 +723,7 @@ _mesa_drawbuffers(struct gl_context *ctx, struct gl_framebuffer *fb,
     * (ex: glDrawBuffer(GL_FRONT_AND_BACK)).
     * Otherwise, destMask[x] can only have one bit set.
     */
-   if (n > 0 && _mesa_bitcount(destMask[0]) > 1) {
+   if (n > 0 && util_bitcount(destMask[0]) > 1) {
       GLuint count = 0, destMask0 = destMask[0];
       while (destMask0) {
          const gl_buffer_index bufIndex = u_bit_scan(&destMask0);
@@ -741,7 +742,7 @@ _mesa_drawbuffers(struct gl_context *ctx, struct gl_framebuffer *fb,
          if (destMask[buf]) {
             gl_buffer_index bufIndex = ffs(destMask[buf]) - 1;
             /* only one bit should be set in the destMask[buf] field */
-            assert(_mesa_bitcount(destMask[buf]) == 1);
+            assert(util_bitcount(destMask[buf]) == 1);
             if (fb->_ColorDrawBufferIndexes[buf] != bufIndex) {
 	       updated_drawbuffers(ctx, fb);
                fb->_ColorDrawBufferIndexes[buf] = bufIndex;

src/mesa/main/imports.c

@@ -213,44 +213,6 @@ _mesa_align_realloc(void *oldBuffer, size_t oldSize, size_t newSize,
 /*@}*/
 
 
-/**********************************************************************/
-/** \name Math */
-/*@{*/
-
-
-#ifndef HAVE___BUILTIN_POPCOUNT
-/**
- * Return number of bits set in given GLuint.
- */
-unsigned int
-_mesa_bitcount(unsigned int n)
-{
-   unsigned int bits;
-   for (bits = 0; n > 0; n = n >> 1) {
-      bits += (n & 1);
-   }
-   return bits;
-}
-#endif
-
-#ifndef HAVE___BUILTIN_POPCOUNTLL
-/**
- * Return number of bits set in given 64-bit uint.
- */
-unsigned int
-_mesa_bitcount_64(uint64_t n)
-{
-   unsigned int bits;
-   for (bits = 0; n > 0; n = n >> 1) {
-      bits += (n & 1);
-   }
-   return bits;
-}
-#endif
-
-/*@}*/
-
-
 /** Needed due to #ifdef's, above. */
 int
 _mesa_vsnprintf(char *str, size_t size, const char *fmt, va_list args)

src/mesa/main/imports.h

@@ -317,21 +317,6 @@ extern void *
 _mesa_align_realloc(void *oldBuffer, size_t oldSize, size_t newSize,
                     unsigned long alignment);
 
-#ifdef HAVE___BUILTIN_POPCOUNT
-#define _mesa_bitcount(i) __builtin_popcount(i)
-#else
-extern unsigned int
-_mesa_bitcount(unsigned int n);
-#endif
-
-#ifdef HAVE___BUILTIN_POPCOUNTLL
-#define _mesa_bitcount_64(i) __builtin_popcountll(i)
-#else
-extern unsigned int
-_mesa_bitcount_64(uint64_t n);
-#endif
-
-
 extern int
 _mesa_snprintf( char *str, size_t size, const char *fmt, ... ) PRINTFLIKE(3, 4);
 

src/mesa/program/program_parse.y

@@ -39,6 +39,8 @@
 #include "program/symbol_table.h"
 #include "program/program_parser.h"
 
+#include "util/u_math.h"
+
 extern void *yy_scan_string(char *);
 extern void yy_delete_buffer(void *);
 
@@ -2607,7 +2609,7 @@ _mesa_parse_arb_program(struct gl_context *ctx, GLenum target, const GLubyte *st
 
    state->prog->arb.NumParameters = state->prog->Parameters->NumParameters;
    state->prog->arb.NumAttributes =
-      _mesa_bitcount_64(state->prog->info.inputs_read);
+      util_bitcount64(state->prog->info.inputs_read);
 
    /*
     * Initialize native counts to logical counts.  The device driver may

src/mesa/state_tracker/st_glsl_to_nir.cpp

@@ -32,7 +32,6 @@
 #include "program/prog_parameter.h"
 #include "program/ir_to_mesa.h"
 #include "main/mtypes.h"
-#include "main/imports.h"
 #include "main/errors.h"
 #include "main/shaderapi.h"
 #include "main/uniforms.h"
@@ -91,10 +90,10 @@ st_nir_assign_vs_in_locations(struct gl_program *prog, nir_shader *nir)
        */
       if (var->data.location == VERT_ATTRIB_EDGEFLAG) {
          /* bit of a hack, mirroring st_translate_vertex_program */
-         var->data.driver_location = _mesa_bitcount_64(nir->info.inputs_read);
+         var->data.driver_location = util_bitcount64(nir->info.inputs_read);
       } else if (nir->info.inputs_read & BITFIELD64_BIT(var->data.location)) {
          var->data.driver_location =
-            _mesa_bitcount_64(nir->info.inputs_read &
+            util_bitcount64(nir->info.inputs_read &
                               BITFIELD64_MASK(var->data.location));
          nir->num_inputs++;
       } else {