Mesa (master): i965: Store the atoms directly in the context

[Kenneth Graunke]

Module: Mesa
Branch: master
Commit: 14aadbe8270845750ce0c62cf511fb5b1bcefd0f
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=14aadbe8270845750ce0c62cf511fb5b1bcefd0f

Author: Ian Romanick <ian.d.romanick at intel.com>
Date:   Tue Nov 11 14:51:29 2014 +0000

i965: Store the atoms directly in the context

Instead of having an extra pointer indirection in one of the hottest
loops in the driver.

On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Batch7:

32-bit: Difference at 95.0% confidence 1.98515% +/- 0.20814% (n=40)
64-bit: Difference at 95.0% confidence 1.5163% +/- 0.811016% (n=60)

v2 (Ken): Cut size of array from 64 to 57 to save memory.

Signed-off-by: Ian Romanick <ian.d.romanick at intel.com>
Reviewed-by: Kenneth Graunke <kenneth at whitecape.org>

---

 src/mesa/drivers/dri/i965/brw_context.h      |    2 +-
 src/mesa/drivers/dri/i965/brw_state_upload.c |   19 ++++++++++++++++---
 2 files changed, 17 insertions(+), 4 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 0e7bd21..a4b29fa 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1383,7 +1383,7 @@ struct brw_context
    } perfmon;
 
    int num_atoms;
-   const struct brw_tracked_state **atoms;
+   const struct brw_tracked_state atoms[57];
 
    /* If (INTEL_DEBUG & DEBUG_BATCH) */
    struct {
diff --git a/src/mesa/drivers/dri/i965/brw_state_upload.c b/src/mesa/drivers/dri/i965/brw_state_upload.c
index 7a25ef5..52d96f4 100644
--- a/src/mesa/drivers/dri/i965/brw_state_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_state_upload.c
@@ -357,6 +357,11 @@ void brw_init_state( struct brw_context *brw )
    const struct brw_tracked_state **atoms;
    int num_atoms;
 
+   STATIC_ASSERT(ARRAY_SIZE(gen4_atoms) <= ARRAY_SIZE(brw->atoms));
+   STATIC_ASSERT(ARRAY_SIZE(gen6_atoms) <= ARRAY_SIZE(brw->atoms));
+   STATIC_ASSERT(ARRAY_SIZE(gen7_atoms) <= ARRAY_SIZE(brw->atoms));
+   STATIC_ASSERT(ARRAY_SIZE(gen8_atoms) <= ARRAY_SIZE(brw->atoms));
+
    brw_init_caches(brw);
 
    if (brw->gen >= 8) {
@@ -373,9 +378,17 @@ void brw_init_state( struct brw_context *brw )
       num_atoms = ARRAY_SIZE(gen4_atoms);
    }
 
-   brw->atoms = atoms;
    brw->num_atoms = num_atoms;
 
+   /* This is to work around brw_context::atoms being declared const.  We want
+    * it to be const, but it needs to be initialized somehow!
+    */
+   struct brw_tracked_state *context_atoms =
+      (struct brw_tracked_state *) &brw->atoms[0];
+
+   for (int i = 0; i < num_atoms; i++)
+      context_atoms[i] = *atoms[i];
+
    while (num_atoms--) {
       assert((*atoms)->dirty.mesa | (*atoms)->dirty.brw);
       assert((*atoms)->emit);
@@ -609,7 +622,7 @@ void brw_upload_state(struct brw_context *brw)
       prev = *state;
 
       for (i = 0; i < brw->num_atoms; i++) {
-	 const struct brw_tracked_state *atom = brw->atoms[i];
+	 const struct brw_tracked_state *atom = &brw->atoms[i];
 	 struct brw_state_flags generated;
 
 	 if (check_state(state, &atom->dirty)) {
@@ -629,7 +642,7 @@ void brw_upload_state(struct brw_context *brw)
    }
    else {
       for (i = 0; i < brw->num_atoms; i++) {
-	 const struct brw_tracked_state *atom = brw->atoms[i];
+	 const struct brw_tracked_state *atom = &brw->atoms[i];
 
 	 if (check_state(state, &atom->dirty)) {
 	    atom->emit(brw);