i965/vs: Improve live interval calculation.

This is derived from the FS visitor code for the same, but tracks each channel
separately (otherwise, some typical fill-a-channel-at-a-time patterns would
produce excessive live intervals across loops and cause spilling).

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48375
          (crash -> failure, can turn into pass by forcing unrolling still)

1 files changed, 306 insertions, 0 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_vec4_live_variables.cpp b/src/mesa/drivers/dri/i965/brw_vec4_live_variables.cpp
new file mode 100644
index 0000000..f34111c
--- /dev/null
+++ b/src/mesa/drivers/dri/i965/brw_vec4_live_variables.cpp
@@ -0,0 +1,306 @@
+/*
+ * Copyright © 2012 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+#include "brw_cfg.h"
+#include "brw_vec4_live_variables.h"
+
+using namespace brw;
+
+/** @file brw_vec4_live_variables.cpp
+ *
+ * Support for computing at the basic block level which variables
+ * (virtual GRFs in our case) are live at entry and exit.
+ *
+ * See Muchnik's Advanced Compiler Design and Implementation, section
+ * 14.1 (p444).
+ */
+
+/**
+ * Sets up the use[] and def[] arrays.
+ *
+ * The basic-block-level live variable analysis needs to know which
+ * variables get used before they're completely defined, and which
+ * variables are completely defined before they're used.
+ *
+ * We independently track each channel of a vec4.  This is because we need to
+ * be able to recognize a sequence like:
+ *
+ * ...
+ * DP4 tmp.x a b;
+ * DP4 tmp.y c d;
+ * MUL result.xy tmp.xy e.xy
+ * ...
+ *
+ * as having tmp live only across that sequence (assuming it's used nowhere
+ * else), because it's a common pattern.  A more conservative approach that
+ * doesn't get tmp marked a deffed in this block will tend to result in
+ * spilling.
+ */
+void
+vec4_live_variables::setup_def_use()
+{
+   int ip = 0;
+
+   for (int b = 0; b < cfg->num_blocks; b++) {
+      bblock_t *block = cfg->blocks[b];
+
+      assert(ip == block->start_ip);
+      if (b > 0)
+	 assert(cfg->blocks[b - 1]->end_ip == ip - 1);
+
+      for (vec4_instruction *inst = (vec4_instruction *)block->start;
+	   inst != block->end->next;
+	   inst = (vec4_instruction *)inst->next) {
+
+	 /* Set use[] for this instruction */
+	 for (unsigned int i = 0; i < 3; i++) {
+	    if (inst->src[i].file == GRF) {
+	       int reg = inst->src[i].reg;
+
+               for (int j = 0; j < 4; j++) {
+                  int c = BRW_GET_SWZ(inst->src[i].swizzle, j);
+                  if (!bd[b].def[reg * 4 + c])
+                     bd[b].use[reg * 4 + c] = true;
+               }
+	    }
+	 }
+
+	 /* Check for unconditional writes to whole registers. These
+	  * are the things that screen off preceding definitions of a
+	  * variable, and thus qualify for being in def[].
+	  */
+	 if (inst->dst.file == GRF &&
+	     v->virtual_grf_sizes[inst->dst.reg] == 1 &&
+	     !inst->predicate) {
+            for (int c = 0; c < 4; c++) {
+               if (inst->dst.writemask & (1 << c)) {
+                  int reg = inst->dst.reg;
+                  if (!bd[b].use[reg * 4 + c])
+                     bd[b].def[reg * 4 + c] = true;
+               }
+            }
+         }
+
+	 ip++;
+      }
+   }
+}
+
+/**
+ * The algorithm incrementally sets bits in liveout and livein,
+ * propagating it through control flow.  It will eventually terminate
+ * because it only ever adds bits, and stops when no bits are added in
+ * a pass.
+ */
+void
+vec4_live_variables::compute_live_variables()
+{
+   bool cont = true;
+
+   while (cont) {
+      cont = false;
+
+      for (int b = 0; b < cfg->num_blocks; b++) {
+	 /* Update livein */
+	 for (int i = 0; i < num_vars; i++) {
+	    if (bd[b].use[i] || (bd[b].liveout[i] && !bd[b].def[i])) {
+	       if (!bd[b].livein[i]) {
+		  bd[b].livein[i] = true;
+		  cont = true;
+	       }
+	    }
+	 }
+
+	 /* Update liveout */
+	 foreach_list(block_node, &cfg->blocks[b]->children) {
+	    bblock_link *link = (bblock_link *)block_node;
+	    bblock_t *block = link->block;
+
+	    for (int i = 0; i < num_vars; i++) {
+	       if (bd[block->block_num].livein[i] && !bd[b].liveout[i]) {
+		  bd[b].liveout[i] = true;
+		  cont = true;
+	       }
+	    }
+	 }
+      }
+   }
+}
+
+vec4_live_variables::vec4_live_variables(vec4_visitor *v, cfg_t *cfg)
+   : v(v), cfg(cfg)
+{
+   mem_ctx = ralloc_context(cfg->mem_ctx);
+
+   num_vars = v->virtual_grf_count * 4;
+   bd = rzalloc_array(mem_ctx, struct block_data, cfg->num_blocks);
+
+   for (int i = 0; i < cfg->num_blocks; i++) {
+      bd[i].def = rzalloc_array(mem_ctx, bool, num_vars);
+      bd[i].use = rzalloc_array(mem_ctx, bool, num_vars);
+      bd[i].livein = rzalloc_array(mem_ctx, bool, num_vars);
+      bd[i].liveout = rzalloc_array(mem_ctx, bool, num_vars);
+   }
+
+   setup_def_use();
+   compute_live_variables();
+}
+
+vec4_live_variables::~vec4_live_variables()
+{
+   ralloc_free(mem_ctx);
+}
+
+#define MAX_INSTRUCTION (1 << 30)
+
+/**
+ * Computes a conservative start/end of the live intervals for each virtual GRF.
+ *
+ * We could expose per-channel live intervals to the consumer based on the
+ * information we computed in vec4_live_variables, except that our only
+ * current user is virtual_grf_interferes().  So we instead union the
+ * per-channel ranges into a per-vgrf range for virtual_grf_def[] and
+ * virtual_grf_use[].
+ *
+ * We could potentially have virtual_grf_interferes() do the test per-channel,
+ * which would let some interesting register allocation occur (particularly on
+ * code-generated GLSL sequences from the Cg compiler which does register
+ * allocation at the GLSL level and thus reuses components of the variable
+ * with distinct lifetimes).  But right now the complexity of doing so doesn't
+ * seem worth it, since having virtual_grf_interferes() be cheap is important
+ * for register allocation performance.
+ */
+void
+vec4_visitor::calculate_live_intervals()
+{
+   if (this->live_intervals_valid)
+      return;
+
+   int *def = ralloc_array(mem_ctx, int, this->virtual_grf_count);
+   int *use = ralloc_array(mem_ctx, int, this->virtual_grf_count);
+   ralloc_free(this->virtual_grf_def);
+   ralloc_free(this->virtual_grf_use);
+   this->virtual_grf_def = def;
+   this->virtual_grf_use = use;
+
+   for (int i = 0; i < this->virtual_grf_count; i++) {
+      def[i] = MAX_INSTRUCTION;
+      use[i] = -1;
+   }
+
+   /* Start by setting up the intervals with no knowledge of control
+    * flow.
+    */
+   int ip = 0;
+   foreach_list(node, &this->instructions) {
+      vec4_instruction *inst = (vec4_instruction *)node;
+
+      for (unsigned int i = 0; i < 3; i++) {
+	 if (inst->src[i].file == GRF) {
+	    int reg = inst->src[i].reg;
+
+	    use[reg] = ip;
+	 }
+      }
+
+      if (inst->dst.file == GRF) {
+         int reg = inst->dst.reg;
+
+         def[reg] = MIN2(def[reg], ip);
+      }
+
+      ip++;
+   }
+
+   /* Now, extend those intervals using our analysis of control flow.
+    *
+    * The control flow-aware analysis was done at a channel level, while at
+    * this point we're distilling it down to vgrfs.
+    */
+   cfg_t cfg(this);
+   vec4_live_variables livevars(this, &cfg);
+
+   for (int b = 0; b < cfg.num_blocks; b++) {
+      for (int i = 0; i < livevars.num_vars; i++) {
+	 if (livevars.bd[b].livein[i]) {
+	    def[i / 4] = MIN2(def[i / 4], cfg.blocks[b]->start_ip);
+	    use[i / 4] = MAX2(use[i / 4], cfg.blocks[b]->start_ip);
+	 }
+
+	 if (livevars.bd[b].liveout[i]) {
+	    def[i / 4] = MIN2(def[i / 4], cfg.blocks[b]->end_ip);
+	    use[i / 4] = MAX2(use[i / 4], cfg.blocks[b]->end_ip);
+	 }
+      }
+   }
+
+   this->live_intervals_valid = true;
+
+   /* Note in the non-control-flow code above, that we only take def[] as the
+    * first store, and use[] as the last use.  We use this in dead code
+    * elimination, to determine when a store never gets used.  However, we
+    * also use these arrays to answer the virtual_grf_interferes() question
+    * (live interval analysis), which is used for register coalescing and
+    * register allocation.
+    *
+    * So, there's a conflict over what the array should mean: if use[]
+    * considers a def after the last use, then the dead code elimination pass
+    * never does anything (and it's an important pass!).  But if we don't
+    * include dead code, then virtual_grf_interferes() lies and we'll do
+    * horrible things like coalesce the register that is dead-code-written
+    * into another register that was live across the dead write (causing the
+    * use of the second register to take the dead write's source value instead
+    * of the coalesced MOV's source value).
+    *
+    * To resolve the conflict, immediately after calculating live intervals,
+    * detect dead code, nuke it, and if we changed anything, calculate again
+    * before returning to the caller.  Now we happen to produce def[] and
+    * use[] arrays that will work for virtual_grf_interferes().
+    */
+   if (dead_code_eliminate())
+      calculate_live_intervals();
+}
+
+bool
+vec4_visitor::virtual_grf_interferes(int a, int b)
+{
+   int a_def = this->virtual_grf_def[a], a_use = this->virtual_grf_use[a];
+   int b_def = this->virtual_grf_def[b], b_use = this->virtual_grf_use[b];
+
+   /* If there's dead code (def but not use), it would break our test
+    * unless we consider it used.
+    */
+   if ((a_use == -1 && a_def != MAX_INSTRUCTION) ||
+       (b_use == -1 && b_def != MAX_INSTRUCTION)) {
+      return true;
+   }
+
+   int start = MAX2(a_def, b_def);
+   int end = MIN2(a_use, b_use);
+
+   return start < end;
+}