radeonsi: initial WIP SI code

This commit adds initial support for acceleration
on SI chips.  egltri is starting to work.

The SI/R600 llvm backend is currently included in mesa
but that may change in the future.

The plan is to write a single gallium driver and
use gallium to support X acceleration.

This commit contains patches from:
Tom Stellard <[email protected]>
Michel Dänzer <[email protected]>
Alex Deucher <[email protected]>
Vadim Girlin <[email protected]>

Signed-off-by: Alex Deucher <[email protected]>

The following commits were squashed in:

======================================================================

radeonsi: Remove unused winsys pointer

This was removed from r600g in commit:

commit 96d882939d612fcc8332f107befec470ed4359de
Author: Marek Olšák <[email protected]>
Date:   Fri Feb 17 01:49:49 2012 +0100

    gallium: remove unused winsys pointers in pipe_screen and pipe_context

    A winsys is already a private object of a driver.

======================================================================

radeonsi: Copy color clamping CAPs from r600

Not sure if the values of these CAPS are correct for radeonsi, but the
same changed were made to r600g in commit:

commit bc1c8369384b5e16547c5bf9728aa78f8dfd66cc
Author: Marek Olšák <[email protected]>
Date:   Mon Jan 23 03:11:17 2012 +0100

    st/mesa: do vertex and fragment color clamping in shaders

    For ARB_color_buffer_float. Most hardware can't do it and st/mesa is
    the perfect place for a fallback.
    The exceptions are:
    - r500 (vertex clamp only)
    - nv50 (both)
    - nvc0 (both)
    - softpipe (both)

    We also have to take into account that r300 can do CLAMPED vertex colors only,
    while r600 can do UNCLAMPED vertex colors only. The difference can be expressed
    with the two new CAPs.

======================================================================

radeonsi: Remove PIPE_CAP_OUTPUT_READ

This CAP was dropped in commit:

commit 04e324008759282728a95a1394bac2c4c2a1a3f9
Author: Marek Olšák <[email protected]>
Date:   Thu Feb 23 23:44:36 2012 +0100

    gallium: remove PIPE_SHADER_CAP_OUTPUT_READ

    r600g is the only driver which has made use of it. The reason the CAP was
    added was to fix some piglit tests when the GLSL pass lower_output_reads
    didn't exist.

    However, not removing output reads breaks the fallback for glClampColorARB,
    which assumes outputs are not readable. The fix would be non-trivial
    and my personal preference is to remove the CAP, considering that reading
    outputs is uncommon and that we can now use lower_output_reads to fix
    the issue that the CAP was supposed to workaround in the first place.

======================================================================

radeonsi: Add missing parameters to rws->buffer_get_tiling() call

This was changed in commit:

commit c0c979eebc076b95cc8d18a013ce2968fe6311ad
Author: Jerome Glisse <[email protected]>
Date:   Mon Jan 30 17:22:13 2012 -0500

    r600g: add support for common surface allocator for tiling v13

    Tiled surface have all kind of alignment constraint that needs to
    be met. Instead of having all this code duplicated btw ddx and
    mesa use common code in libdrm_radeon this also ensure that both
    ddx and mesa compute those alignment in the same way.

    v2 fix evergreen
    v3 fix compressed texture and workaround cube texture issue by
       disabling 2D array mode for cubemap (need to check if r7xx and
       newer are also affected by the issue)
    v4 fix texture array
    v5 fix evergreen and newer, split surface values computation from
       mipmap tree generation so that we can get them directly from the
       ddx
    v6 final fix to evergreen tile split value
    v7 fix mipmap offset to avoid to use random value, use color view
       depth view to address different layer as hardware is doing some
       magic rotation depending on the layer
    v8 fix COLOR_VIEW on r6xx for linear array mode, use COLOR_VIEW on
       evergreen, align bytes per pixel to a multiple of a dword
    v9 fix handling of stencil on evergreen, half fix for compressed
       texture
    v10 fix evergreen compressed texture proper support for stencil
        tile split. Fix stencil issue when array mode was clear by
        the kernel, always program stencil bo. On evergreen depth
        buffer bo need to be big enough to hold depth buffer + stencil
        buffer as even with stencil disabled things get written there.
    v11 rebase on top of mesa, fix pitch issue with 1d surface on evergreen,
        old ddx overestimate those. Fix linear case when pitch*height < 64.
        Fix r300g.
    v12 Fix linear case when pitch*height < 64 for old path, adapt to
        libdrm API change
    v13 add libdrm check

    Signed-off-by: Jerome Glisse <[email protected]>

======================================================================

radeonsi: Remove PIPE_TRANSFER_MAP_PERMANENTLY

This was removed in commit:

commit 62f44f670bb0162e89fd4786af877f8da9ff607c
Author: Marek Olšák <[email protected]>
Date:   Mon Mar 5 13:45:00 2012 +0100

    Revert "gallium: add flag PIPE_TRANSFER_MAP_PERMANENTLY"

    This reverts commit 0950086376b1c8b7fb89eda81ed7f2f06dee58bc.

    It was decided to refactor the transfer API instead of adding workarounds
    to address the performance issues.

======================================================================

radeonsi: Handle PIPE_VIDEO_CAP_PREFERED_FORMAT.

Reintroduced in commit 9d9afcb5bac2931d4b8e6d1aa571e941c5110c90.

======================================================================

radeonsi: nuke the fallback for vertex and fragment color clamping

Ported from r600g commit c2b800cf38b299c1ab1c53dc0e4ea00c7acef853.

======================================================================

radeonsi: don't expose transform_feedback2 without kernel support

Ported from r600g commit 15146fd1bcbb08e44a1cbb984440ee1a5de63d48.

======================================================================

radeonsi: Handle PIPE_CAP_GLSL_FEATURE_LEVEL.

Ported from r600g part of commit 171be755223d99f8cc5cc1bdaf8bd7b4caa04b4f.

======================================================================

radeonsi: set minimum point size to 1.0 for non-sprite non-aa points.

Ported from r600g commit f183cc9ce3ad1d043bdf8b38fd519e8f437714fc.

======================================================================

radeonsi: rework and consolidate stencilref state setting.

Ported from r600g commit a2361946e782b57f0c63587841ca41c0ea707070.

======================================================================

radeonsi: cleanup setting DB_SHADER_CONTROL.

Ported from r600g commit 3d061caaed13b646ff40754f8ebe73f3d4983c5b.

======================================================================

radeonsi: Get rid of register masks.

Ported from r600g commits
3d061caaed13b646ff40754f8ebe73f3d4983c5b..9344ab382a1765c1a7c2560e771485edf4954fe2.

======================================================================

radeonsi: get rid of r600_context_reg.

Ported from r600g commits
9344ab382a1765c1a7c2560e771485edf4954fe2..bed20f02a771f43e1c5092254705701c228cfa7f.

======================================================================

radeonsi: Fix regression from 'Get rid of register masks'.

======================================================================

radeonsi: optimize r600_resource_va.

Ported from r600g commit 669d8766ff3403938794eb80d7769347b6e52174.

======================================================================

radeonsi: remove u8,u16,u32,u64 types.

Ported from r600g commit 78293b99b23268e6698f1267aaf40647c17d95a5.

======================================================================

radeonsi: merge r600_context with r600_pipe_context.

Ported from r600g commit e4340c1908a6a3b09e1a15d5195f6da7d00494d0.

======================================================================

radeonsi: Miscellaneous context cleanups.

Ported from r600g commits
e4340c1908a6a3b09e1a15d5195f6da7d00494d0..621e0db71c5ddcb379171064a4f720c9cf01e888.

======================================================================

radeonsi: add a new simple API for state emission.

Ported from r600g commits
621e0db71c5ddcb379171064a4f720c9cf01e888..f661405637bba32c2cfbeecf6e2e56e414e9521e.

======================================================================

radeonsi: Also remove sbu_flags member of struct r600_reg.

Requires using sid.h instead of r600d.h for the new CP_COHER_CNTL definitions,
so some code needs to be disabled for now.

======================================================================

radeonsi: Miscellaneous simplifications.

Ported from r600g commits 38bf2763482b4f1b6d95cd51aecec75601d8b90f and
b0337b679ad4c2feae59215104cfa60b58a619d5.

======================================================================

radeonsi: Handle PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION.

Ported from commit 8b4f7b0672d663273310fffa9490ad996f5b914a.

======================================================================

radeonsi: Use a fake reloc to sleep for fences.

Ported from r600g commit 8cd03b933cf868ff867e2db4a0937005a02fd0e4.

======================================================================

radeonsi: adapt to get_query_result interface change.

Ported from r600g commit 4445e170bee23a3607ece0e010adef7058ac6a11.

1 files changed, 362 insertions, 0 deletions

diff --git a/src/gallium/drivers/radeon/AMDILUtilityFunctions.h b/src/gallium/drivers/radeon/AMDILUtilityFunctions.h
new file mode 100644
index 00000000000..637c868b55c
--- /dev/null
+++ b/src/gallium/drivers/radeon/AMDILUtilityFunctions.h
@@ -0,0 +1,362 @@
+//===-- AMDILUtilityFunctions.h - AMDIL Utility Functions Header --------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//==-----------------------------------------------------------------------===//
+//
+// This file provides declarations for functions that are used across different
+// classes and provide various conversions or utility to shorten the code
+//
+//===----------------------------------------------------------------------===//
+#ifndef AMDILUTILITYFUNCTIONS_H_
+#define AMDILUTILITYFUNCTIONS_H_
+
+#include "AMDIL.h"
+#include "AMDILTargetMachine.h"
+#include "llvm/ADT/SmallVector.h"
+
+// Utility functions from ID
+//
+namespace llvm {
+class TargetRegisterClass;
+class SDValue;
+class SDNode;
+class Value;
+class Type;
+class StructType;
+class IntegerType;
+class FunctionType;
+class VectorType;
+class ArrayType;
+class PointerType;
+class OpaqueType;
+class MachineInstr;
+
+}
+enum SrcSwizzles {
+  AMDIL_SRC_SWIZZLE_DEFAULT = 0,
+  AMDIL_SRC_SWIZZLE_X000,
+  AMDIL_SRC_SWIZZLE_0X00,
+  AMDIL_SRC_SWIZZLE_00X0,
+  AMDIL_SRC_SWIZZLE_000X,
+  AMDIL_SRC_SWIZZLE_Y000,
+  AMDIL_SRC_SWIZZLE_0Y00,
+  AMDIL_SRC_SWIZZLE_00Y0,
+  AMDIL_SRC_SWIZZLE_000Y,
+  AMDIL_SRC_SWIZZLE_Z000,
+  AMDIL_SRC_SWIZZLE_0Z00,
+  AMDIL_SRC_SWIZZLE_00Z0,
+  AMDIL_SRC_SWIZZLE_000Z,
+  AMDIL_SRC_SWIZZLE_W000,
+  AMDIL_SRC_SWIZZLE_0W00,
+  AMDIL_SRC_SWIZZLE_00W0,
+  AMDIL_SRC_SWIZZLE_000W,
+  AMDIL_SRC_SWIZZLE_XY00,
+  AMDIL_SRC_SWIZZLE_00XY,
+  AMDIL_SRC_SWIZZLE_ZW00,
+  AMDIL_SRC_SWIZZLE_00ZW,
+  AMDIL_SRC_SWIZZLE_XYZ0,
+  AMDIL_SRC_SWIZZLE_0XYZ,
+  AMDIL_SRC_SWIZZLE_XYZW,
+  AMDIL_SRC_SWIZZLE_0000,
+  AMDIL_SRC_SWIZZLE_XXXX,
+  AMDIL_SRC_SWIZZLE_YYYY,
+  AMDIL_SRC_SWIZZLE_ZZZZ,
+  AMDIL_SRC_SWIZZLE_WWWW,
+  AMDIL_SRC_SWIZZLE_XYXY,
+  AMDIL_SRC_SWIZZLE_ZWZW,
+  AMDIL_SRC_SWIZZLE_XZXZ,
+  AMDIL_SRC_SWIZZLE_YWYW,
+  AMDIL_SRC_SWIZZLE_X0Y0,
+  AMDIL_SRC_SWIZZLE_0X0Y,
+  AMDIL_SRC_SWIZZLE_XY_NEGY,
+  AMDIL_SRC_SWIZZLE_NEGYW,
+  AMDIL_SRC_SWIZZLE_NEGX,
+  AMDIL_SRC_SWIZZLE_XY_NEGXY,
+  AMDIL_SRC_SWIZZLE_NEG_XYZW,
+  AMDIL_SRC_SWIZZLE_0YZW,
+  AMDIL_SRC_SWIZZLE_X0ZW,
+  AMDIL_SRC_SWIZZLE_XY0W,
+  AMDIL_SRC_SWIZZLE_X,
+  AMDIL_SRC_SWIZZLE_Y,
+  AMDIL_SRC_SWIZZLE_Z,
+  AMDIL_SRC_SWIZZLE_W,
+  AMDIL_SRC_SWIZZLE_XY,
+  AMDIL_SRC_SWIZZLE_ZW,
+  AMDIL_SRC_SWIZZLE_LAST
+};
+enum DstSwizzles {
+  AMDIL_DST_SWIZZLE_DEFAULT = 0,
+  AMDIL_DST_SWIZZLE_X___,
+  AMDIL_DST_SWIZZLE_XY__,
+  AMDIL_DST_SWIZZLE_XYZ_,
+  AMDIL_DST_SWIZZLE_XYZW,
+  AMDIL_DST_SWIZZLE__Y__,
+  AMDIL_DST_SWIZZLE__YZ_,
+  AMDIL_DST_SWIZZLE__YZW,
+  AMDIL_DST_SWIZZLE___Z_,
+  AMDIL_DST_SWIZZLE___ZW,
+  AMDIL_DST_SWIZZLE____W,
+  AMDIL_DST_SWIZZLE_X_ZW,
+  AMDIL_DST_SWIZZLE_XY_W,
+  AMDIL_DST_SWIZZLE_X_Z_,
+  AMDIL_DST_SWIZZLE_X__W,
+  AMDIL_DST_SWIZZLE__Y_W,
+  AMDIL_DST_SWIZZLE_LAST
+};
+// Function to get the correct src swizzle string from ID
+const char *getSrcSwizzle(unsigned);
+
+// Function to get the correct dst swizzle string from ID
+const char *getDstSwizzle(unsigned);
+
+const llvm::TargetRegisterClass *getRegClassFromID(unsigned int ID);
+
+unsigned int getMoveInstFromID(unsigned int ID);
+unsigned int getPHIMoveInstFromID(unsigned int ID);
+
+// Utility functions from Type.
+const llvm::TargetRegisterClass *getRegClassFromType(unsigned int type);
+unsigned int getTargetIndependentMoveFromType(unsigned int type);
+
+// Debug functions for SDNode and SDValue.
+void printSDValue(const llvm::SDValue &Op, int level);
+void printSDNode(const llvm::SDNode *N);
+
+// Functions to check if an opcode is a specific type.
+bool isMove(unsigned int opcode);
+bool isPHIMove(unsigned int opcode);
+bool isMoveOrEquivalent(unsigned int opcode);
+
+// Function to check address space
+bool check_type(const llvm::Value *ptr, unsigned int addrspace);
+
+// Group of functions that recursively calculate the size of a structure based
+// on it's sub-types.
+size_t getTypeSize(llvm::Type * const T, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::StructType * const ST, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::IntegerType * const IT, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::FunctionType * const FT, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::ArrayType * const AT, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::VectorType * const VT, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::PointerType * const PT, bool dereferencePtr = false);
+size_t
+getTypeSize(llvm::OpaqueType * const OT, bool dereferencePtr = false);
+
+// Group of functions that recursively calculate the number of elements of a
+// structure based on it's sub-types.
+size_t getNumElements(llvm::Type * const T);
+size_t getNumElements(llvm::StructType * const ST);
+size_t getNumElements(llvm::IntegerType * const IT);
+size_t getNumElements(llvm::FunctionType * const FT);
+size_t getNumElements(llvm::ArrayType * const AT);
+size_t getNumElements(llvm::VectorType * const VT);
+size_t getNumElements(llvm::PointerType * const PT);
+size_t getNumElements(llvm::OpaqueType * const OT);
+const llvm::Value *getBasePointerValue(const llvm::Value *V);
+const llvm::Value *getBasePointerValue(const llvm::MachineInstr *MI);
+
+
+int64_t GET_SCALAR_SIZE(llvm::Type* A);
+
+// Helper functions that check the opcode for status information
+bool isLoadInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isExtLoadInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isSWSExtLoadInst(llvm::MachineInstr *MI);
+bool isSExtLoadInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isZExtLoadInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isAExtLoadInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isStoreInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isTruncStoreInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isAtomicInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isVolatileInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isGlobalInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isPrivateInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isConstantInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isRegionInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isLocalInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isImageInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isAppendInst(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isRegionAtomic(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isLocalAtomic(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isGlobalAtomic(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+bool isArenaAtomic(const llvm::TargetInstrInfo * TII, llvm::MachineInstr *MI);
+
+
+// Macros that are used to help with switch statements for various data types
+// However, these macro's do not return anything unlike the second set below.
+#define ExpandCaseTo32bitIntTypes(Instr)  \
+case Instr##_i8: \
+case Instr##_i16: \
+case Instr##_i32:
+
+#define ExpandCaseTo32bitIntTruncTypes(Instr)  \
+case Instr##_i16i8: \
+case Instr##_i32i8: \
+case Instr##_i32i16: 
+
+#define ExpandCaseToIntTypes(Instr) \
+    ExpandCaseTo32bitIntTypes(Instr) \
+case Instr##_i64:
+
+#define ExpandCaseToIntTruncTypes(Instr) \
+    ExpandCaseTo32bitIntTruncTypes(Instr) \
+case Instr##_i64i8:\
+case Instr##_i64i16:\
+case Instr##_i64i32:\
+
+#define ExpandCaseToFloatTypes(Instr) \
+    case Instr##_f32: \
+case Instr##_f64:
+
+#define ExpandCaseToFloatTruncTypes(Instr) \
+case Instr##_f64f32:
+
+#define ExpandCaseTo32bitScalarTypes(Instr) \
+    ExpandCaseTo32bitIntTypes(Instr) \
+case Instr##_f32:
+
+#define ExpandCaseToAllScalarTypes(Instr) \
+    ExpandCaseToFloatTypes(Instr) \
+ExpandCaseToIntTypes(Instr)
+
+#define ExpandCaseToAllScalarTruncTypes(Instr) \
+    ExpandCaseToFloatTruncTypes(Instr) \
+ExpandCaseToIntTruncTypes(Instr)
+
+// Vector versions of above macros
+#define ExpandCaseToVectorIntTypes(Instr) \
+    case Instr##_v2i8: \
+case Instr##_v4i8: \
+case Instr##_v2i16: \
+case Instr##_v4i16: \
+case Instr##_v2i32: \
+case Instr##_v4i32: \
+case Instr##_v2i64:
+
+#define ExpandCaseToVectorIntTruncTypes(Instr) \
+case Instr##_v2i16i8: \
+case Instr##_v4i16i8: \
+case Instr##_v2i32i8: \
+case Instr##_v4i32i8: \
+case Instr##_v2i32i16: \
+case Instr##_v4i32i16: \
+case Instr##_v2i64i8: \
+case Instr##_v2i64i16: \
+case Instr##_v2i64i32: 
+
+#define ExpandCaseToVectorFloatTypes(Instr) \
+    case Instr##_v2f32: \
+case Instr##_v4f32: \
+case Instr##_v2f64:
+
+#define ExpandCaseToVectorFloatTruncTypes(Instr) \
+case Instr##_v2f64f32:
+
+#define ExpandCaseToVectorByteTypes(Instr) \
+  case Instr##_v4i8:\
+case Instr##_v2i16: \
+case Instr##_v4i16:
+
+#define ExpandCaseToAllVectorTypes(Instr) \
+    ExpandCaseToVectorFloatTypes(Instr) \
+ExpandCaseToVectorIntTypes(Instr)
+
+#define ExpandCaseToAllVectorTruncTypes(Instr) \
+    ExpandCaseToVectorFloatTruncTypes(Instr) \
+ExpandCaseToVectorIntTruncTypes(Instr)
+
+#define ExpandCaseToAllTypes(Instr) \
+    ExpandCaseToAllVectorTypes(Instr) \
+ExpandCaseToAllScalarTypes(Instr)
+
+#define ExpandCaseToAllTruncTypes(Instr) \
+    ExpandCaseToAllVectorTruncTypes(Instr) \
+ExpandCaseToAllScalarTruncTypes(Instr)
+
+#define ExpandCaseToPackedTypes(Instr) \
+    case Instr##_v2i8: \
+    case Instr##_v4i8: \
+    case Instr##_v2i16: \
+    case Instr##_v4i16:
+
+#define ExpandCaseToByteShortTypes(Instr) \
+    case Instr##_i8: \
+    case Instr##_i16: \
+    ExpandCaseToPackedTypes(Instr)
+
+// Macros that expand into case statements with return values
+#define ExpandCaseTo32bitIntReturn(Instr, Return)  \
+case Instr##_i8: return Return##_i8;\
+case Instr##_i16: return Return##_i16;\
+case Instr##_i32: return Return##_i32;
+
+#define ExpandCaseToIntReturn(Instr, Return) \
+    ExpandCaseTo32bitIntReturn(Instr, Return) \
+case Instr##_i64: return Return##_i64;
+
+#define ExpandCaseToFloatReturn(Instr, Return) \
+    case Instr##_f32: return Return##_f32;\
+case Instr##_f64: return Return##_f64;
+
+#define ExpandCaseToAllScalarReturn(Instr, Return) \
+    ExpandCaseToFloatReturn(Instr, Return) \
+ExpandCaseToIntReturn(Instr, Return)
+
+// These macros expand to common groupings of RegClass ID's
+#define ExpandCaseTo1CompRegID \
+case AMDIL::GPRI8RegClassID: \
+case AMDIL::GPRI16RegClassID: \
+case AMDIL::GPRI32RegClassID: \
+case AMDIL::GPRF32RegClassID:
+
+#define ExpandCaseTo2CompRegID \
+    case AMDIL::GPRI64RegClassID: \
+case AMDIL::GPRF64RegClassID: \
+case AMDIL::GPRV2I8RegClassID: \
+case AMDIL::GPRV2I16RegClassID: \
+case AMDIL::GPRV2I32RegClassID: \
+case AMDIL::GPRV2F32RegClassID:
+
+// Macros that expand to case statements for specific bitlengths
+#define ExpandCaseTo8BitType(Instr) \
+    case Instr##_i8:
+
+#define ExpandCaseTo16BitType(Instr) \
+    case Instr##_v2i8: \
+case Instr##_i16:
+
+#define ExpandCaseTo32BitType(Instr) \
+    case Instr##_v4i8: \
+case Instr##_v2i16: \
+case Instr##_i32: \
+case Instr##_f32:
+
+#define ExpandCaseTo64BitType(Instr) \
+    case Instr##_v4i16: \
+case Instr##_v2i32: \
+case Instr##_v2f32: \
+case Instr##_i64: \
+case Instr##_f64:
+
+#define ExpandCaseTo128BitType(Instr) \
+    case Instr##_v4i32: \
+case Instr##_v4f32: \
+case Instr##_v2i64: \
+case Instr##_v2f64:
+
+bool commaPrint(int i, llvm::raw_ostream &O);
+/// Helper function to get the currently get/set flags.
+void getAsmPrinterFlags(llvm::MachineInstr *MI, llvm::AMDILAS::InstrResEnc &curRes);
+void setAsmPrinterFlags(llvm::MachineInstr *MI, llvm::AMDILAS::InstrResEnc &curRes);
+
+#endif // AMDILUTILITYFUNCTIONS_H_