commit e4d9ab609f217bde95722e76a98d646890c9abc9
author Iago Toral Quiroga <itoral@igalia.com> Wed Feb 17 10:05:02 2016 +0100
committer Samuel Iglesias Gonsálvez <siglesias@igalia.com> Tue Jan 03 11:26:51 2017 +0100
tree b90bfacc9ee2f2630d2767b46ed8bab880244423
parent 07bc6a35d3d6d94d45b81bd10002f0e420d855c2

i965/vec4: fix scratch offset for 64bit data

A vec4 is 16 bytes and a dvec4 is 32 bytes so for doubles we have
to multiply the reladdr by 2. The reg_offset part is in units of 16
bytes and is used to select the low/high 16-byte chunk of a full
dvec4, so we don't want to multiply that part of the address.

Reviewed-by: Matt Turner <mattst88@gmail.com>

src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp

diff --git a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
index 9a98906..d4fe4fa 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
@@ -1443,13 +1443,23 @@
       message_header_scale *= 16;
 
    if (reladdr) {
+      /* A vec4 is 16 bytes and a dvec4 is 32 bytes so for doubles we have
+       * to multiply the reladdr by 2. Notice that the reg_offset part
+       * is in units of 16 bytes and is used to select the low/high 16-byte
+       * chunk of a full dvec4, so we don't want to multiply that part.
+       */
       src_reg index = src_reg(this, glsl_type::int_type);
-
-      emit_before(block, inst, ADD(dst_reg(index), *reladdr,
-                                   brw_imm_d(reg_offset)));
-      emit_before(block, inst, MUL(dst_reg(index), index,
-                                   brw_imm_d(message_header_scale)));
-
+      if (type_sz(inst->dst.type) < 8) {
+         emit_before(block, inst, ADD(dst_reg(index), *reladdr,
+                                      brw_imm_d(reg_offset)));
+         emit_before(block, inst, MUL(dst_reg(index), index,
+                                      brw_imm_d(message_header_scale)));
+      } else {
+         emit_before(block, inst, MUL(dst_reg(index), *reladdr,
+                                      brw_imm_d(message_header_scale * 2)));
+         emit_before(block, inst, ADD(dst_reg(index), index,
+                                      brw_imm_d(reg_offset * message_header_scale)));
+      }
       return index;
    } else {
       return brw_imm_d(reg_offset * message_header_scale);