i965: Introduce new SHADER_OPCODE_URB_WRITE_SIMD8_MASKED/PER_SLOT opcodes.

In the vec4 backend, we have a vec4_instruction::urb_write_flags field.
There are many kinds of flags for SIMD4x2 messages.

However, there are really only two (per-slot offset, use channel masks)
for SIMD8 messages.  Rather than adding a boolean flag for per-slot
offsets (polluting all instructions), I decided to just make three new
opcodes.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>

src/mesa/drivers/dri/i965/brw_defines.h

    SHADER_OPCODE_GEN7_SCRATCH_READ,
 
    SHADER_OPCODE_URB_WRITE_SIMD8,
+   SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT,
+   SHADER_OPCODE_URB_WRITE_SIMD8_MASKED,
+   SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT,
 
    /**
     * Return the index of an arbitrary live channel (i.e. one of the channels

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

    case SHADER_OPCODE_TYPED_SURFACE_READ:
    case SHADER_OPCODE_TYPED_SURFACE_WRITE:
    case SHADER_OPCODE_URB_WRITE_SIMD8:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT:
       return true;
    case FS_OPCODE_UNIFORM_PULL_CONSTANT_LOAD:
       return src[1].file == GRF;
    switch (opcode) {
    case FS_OPCODE_FB_WRITE:
    case SHADER_OPCODE_URB_WRITE_SIMD8:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT:
    case SHADER_OPCODE_UNTYPED_ATOMIC:
    case SHADER_OPCODE_UNTYPED_SURFACE_READ:
    case SHADER_OPCODE_UNTYPED_SURFACE_WRITE:
    case SHADER_OPCODE_TYPED_SURFACE_READ:
    case SHADER_OPCODE_TYPED_SURFACE_WRITE:
    case SHADER_OPCODE_URB_WRITE_SIMD8:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT:
    case FS_OPCODE_INTERPOLATE_AT_CENTROID:
    case FS_OPCODE_INTERPOLATE_AT_SAMPLE:
    case FS_OPCODE_INTERPOLATE_AT_SHARED_OFFSET:

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

    brw_inst_set_sfid(p->devinfo, insn, BRW_SFID_URB);
    brw_inst_set_urb_opcode(p->devinfo, insn, GEN8_URB_OPCODE_SIMD8_WRITE);
 
+   if (inst->opcode == SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT ||
+       inst->opcode == SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT)
+      brw_inst_set_urb_per_slot_offset(p->devinfo, insn, true);
+
+   if (inst->opcode == SHADER_OPCODE_URB_WRITE_SIMD8_MASKED ||
+       inst->opcode == SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT)
+      brw_inst_set_urb_channel_mask_present(p->devinfo, insn, true);
+
    brw_inst_set_mlen(p->devinfo, insn, inst->mlen);
    brw_inst_set_rlen(p->devinfo, insn, 0);
    brw_inst_set_eot(p->devinfo, insn, inst->eot);
 	 break;
 
       case SHADER_OPCODE_URB_WRITE_SIMD8:
+      case SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT:
+      case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED:
+      case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT:
 	 generate_urb_write(inst, src[0]);
 	 break;
 

src/mesa/drivers/dri/i965/brw_inst.h

    /* 4-6: */ -1, -1, -1, -1, -1, -1, -1, -1,
    /* 7:   */ MD(16), MD(16),
    /* 8:   */ MD(17), MD(17))
+FC(urb_channel_mask_present, MD(15), MD(15), devinfo->gen >= 8)
 FC(urb_complete, MD(15), MD(15), devinfo->gen < 8)
 FC(urb_used, MD(14), MD(14), devinfo->gen < 7)
 FC(urb_allocate, MD(13), MD(13), devinfo->gen < 7)

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

       return "gen7_scratch_read";
    case SHADER_OPCODE_URB_WRITE_SIMD8:
       return "gen8_urb_write_simd8";
+   case SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT:
+      return "gen8_urb_write_simd8_per_slot";
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED:
+      return "gen8_urb_write_simd8_masked";
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT:
+      return "gen8_urb_write_simd8_masked_per_slot";
 
    case SHADER_OPCODE_FIND_LIVE_CHANNEL:
       return "find_live_channel";
    case SHADER_OPCODE_TYPED_SURFACE_WRITE_LOGICAL:
    case SHADER_OPCODE_MEMORY_FENCE:
    case SHADER_OPCODE_URB_WRITE_SIMD8:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_PER_SLOT:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED:
+   case SHADER_OPCODE_URB_WRITE_SIMD8_MASKED_PER_SLOT:
    case FS_OPCODE_FB_WRITE:
    case SHADER_OPCODE_BARRIER:
       return true;