Mesa (master): intel/fs: Add explicit last_rt flag to fb writes orthogonal to eot.

[Jason Ekstrand]

Module: Mesa
Branch: master
Commit: 4bd2047deea31e877ae023a3845f925aab69cdc7
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=4bd2047deea31e877ae023a3845f925aab69cdc7

Author: Francisco Jerez <currojerez at riseup.net>
Date:   Fri Jan 13 14:01:45 2017 -0800

intel/fs: Add explicit last_rt flag to fb writes orthogonal to eot.

When using multiple RT write messages to the same RT such as for
dual-source blending or all RT writes in SIMD32, we have to set the
"Last Render Target Select" bit on all write messages that target the
last RT but only set EOT on the last RT write in the shader.
Special-casing for dual-source blend works today because that is the
only case which requires multiple RT write messages per RT.  When we
start doing SIMD32, this will become much more common so we add a
dedicated bit for it.

Reviewed-by: Jason Ekstrand <jason at jlekstrand.net>
Reviewed-by: Matt Turner <mattst88 at gmail.com>

---

 src/intel/compiler/brw_fs.cpp           | 1 +
 src/intel/compiler/brw_fs_generator.cpp | 6 +-----
 src/intel/compiler/brw_fs_visitor.cpp   | 2 ++
 src/intel/compiler/brw_ir_fs.h          | 1 +
 4 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/src/intel/compiler/brw_fs.cpp b/src/intel/compiler/brw_fs.cpp
index 8d9278684f..d67c0a4192 100644
--- a/src/intel/compiler/brw_fs.cpp
+++ b/src/intel/compiler/brw_fs.cpp
@@ -3246,6 +3246,7 @@ fs_visitor::emit_repclear_shader()
       }
    }
    write->eot = true;
+   write->last_rt = true;
 
    calculate_cfg();
 
diff --git a/src/intel/compiler/brw_fs_generator.cpp b/src/intel/compiler/brw_fs_generator.cpp
index f310a84e25..f49ab442fb 100644
--- a/src/intel/compiler/brw_fs_generator.cpp
+++ b/src/intel/compiler/brw_fs_generator.cpp
@@ -289,10 +289,6 @@ fs_generator::fire_fb_write(fs_inst *inst,
     */
    const uint32_t surf_index = inst->target;
 
-   bool last_render_target = inst->eot ||
-                             (prog_data->dual_src_blend && dispatch_width == 16);
-
-
    brw_fb_WRITE(p,
                 payload,
                 implied_header,
@@ -301,7 +297,7 @@ fs_generator::fire_fb_write(fs_inst *inst,
                 nr,
                 0,
                 inst->eot,
-                last_render_target,
+                inst->last_rt,
                 inst->header_size != 0);
 
    brw_mark_surface_used(&prog_data->base, surf_index);
diff --git a/src/intel/compiler/brw_fs_visitor.cpp b/src/intel/compiler/brw_fs_visitor.cpp
index 41dbd76106..a24808eac6 100644
--- a/src/intel/compiler/brw_fs_visitor.cpp
+++ b/src/intel/compiler/brw_fs_visitor.cpp
@@ -103,6 +103,7 @@ fs_visitor::emit_dummy_fs()
    fs_inst *write;
    write = bld.emit(FS_OPCODE_FB_WRITE);
    write->eot = true;
+   write->last_rt = true;
    if (devinfo->gen >= 6) {
       write->base_mrf = 2;
       write->mlen = 4 * reg_width;
@@ -459,6 +460,7 @@ fs_visitor::emit_fb_writes()
       inst->target = 0;
    }
 
+   inst->last_rt = true;
    inst->eot = true;
 }
 
diff --git a/src/intel/compiler/brw_ir_fs.h b/src/intel/compiler/brw_ir_fs.h
index f06a33c516..92dad269a3 100644
--- a/src/intel/compiler/brw_ir_fs.h
+++ b/src/intel/compiler/brw_ir_fs.h
@@ -374,6 +374,7 @@ public:
 
    uint8_t sources; /**< Number of fs_reg sources. */
 
+   bool last_rt:1;
    bool pi_noperspective:1;   /**< Pixel interpolator noperspective flag */
 };