[PATCH] D13957: AMDGPU: Define correct number of SGPRs

[Matt Arsenault via llvm-commits]

arsenm created this revision.
arsenm added a reviewer: tstellarAMD.
arsenm added a subscriber: llvm-commits.
Herald added a subscriber: arsenm.

There are actually 104 so 2 were missing.
    
More assembler tests with high register number tuples
will be included in later patches.

http://reviews.llvm.org/D13957

Files:
  lib/Target/AMDGPU/SIRegisterInfo.cpp
  lib/Target/AMDGPU/SIRegisterInfo.td
  test/MC/AMDGPU/sop2.s

Index: test/MC/AMDGPU/sop2.s
===================================================================
--- test/MC/AMDGPU/sop2.s
+++ test/MC/AMDGPU/sop2.s
@@ -129,3 +129,6 @@
 
 // CHECK: s_absdiff_i32 s2, s4, s6 ; encoding: [0x04,0x06,0x02,0x96]
 s_absdiff_i32 s2, s4, s6
+
+// CHECK: s_add_u32 s101, s102, s103 ; encoding: [0x66,0x67,0x65,0x80]
+s_add_u32 s101, s102, s103
Index: lib/Target/AMDGPU/SIRegisterInfo.td
===================================================================
--- lib/Target/AMDGPU/SIRegisterInfo.td
+++ lib/Target/AMDGPU/SIRegisterInfo.td
@@ -56,7 +56,7 @@
 }
 
 // SGPR registers
-foreach Index = 0-101 in {
+foreach Index = 0-103 in {
   def SGPR#Index : SIReg <"SGPR"#Index, Index>;
 }
 
@@ -75,23 +75,23 @@
 
 // SGPR 32-bit registers
 def SGPR_32 : RegisterClass<"AMDGPU", [i32, f32], 32,
-                            (add (sequence "SGPR%u", 0, 101))>;
+                            (add (sequence "SGPR%u", 0, 103))>;
 
 // SGPR 64-bit registers
 def SGPR_64Regs : RegisterTuples<[sub0, sub1],
-                             [(add (decimate (trunc SGPR_32, 101), 2)),
+                             [(add (decimate SGPR_32, 2)),
                               (add (decimate (shl SGPR_32, 1), 2))]>;
 
 // SGPR 128-bit registers
 def SGPR_128 : RegisterTuples<[sub0, sub1, sub2, sub3],
-                              [(add (decimate (trunc SGPR_32, 99), 4)),
+                              [(add (decimate SGPR_32, 4)),
                                (add (decimate (shl SGPR_32, 1), 4)),
                                (add (decimate (shl SGPR_32, 2), 4)),
                                (add (decimate (shl SGPR_32, 3), 4))]>;
 
 // SGPR 256-bit registers
 def SGPR_256 : RegisterTuples<[sub0, sub1, sub2, sub3, sub4, sub5, sub6, sub7],
-                              [(add (decimate (trunc SGPR_32, 95), 4)),
+                              [(add (decimate SGPR_32, 4)),
                                (add (decimate (shl SGPR_32, 1), 4)),
                                (add (decimate (shl SGPR_32, 2), 4)),
                                (add (decimate (shl SGPR_32, 3), 4)),
@@ -103,7 +103,7 @@
 // SGPR 512-bit registers
 def SGPR_512 : RegisterTuples<[sub0, sub1, sub2, sub3, sub4, sub5, sub6, sub7,
                                sub8, sub9, sub10, sub11, sub12, sub13, sub14, sub15],
-                              [(add (decimate (trunc SGPR_32, 87), 4)),
+                              [(add (decimate SGPR_32, 4)),
                                (add (decimate (shl SGPR_32, 1), 4)),
                                (add (decimate (shl SGPR_32, 2), 4)),
                                (add (decimate (shl SGPR_32, 3), 4)),
Index: lib/Target/AMDGPU/SIRegisterInfo.cpp
===================================================================
--- lib/Target/AMDGPU/SIRegisterInfo.cpp
+++ lib/Target/AMDGPU/SIRegisterInfo.cpp
@@ -41,6 +41,10 @@
   reserveRegisterTuples(Reserved, AMDGPU::EXEC);
   reserveRegisterTuples(Reserved, AMDGPU::FLAT_SCR);
 
+  // Reserve the last 2 registers so we will always have at least 2 more that
+  // will physically contain VCC.
+  reserveRegisterTuples(Reserved, AMDGPU::SGPR102_SGPR103);
+
   // Tonga and Iceland can only allocate a fixed number of SGPRs due
   // to a hw bug.
   if (MF.getSubtarget<AMDGPUSubtarget>().hasSGPRInitBug()) {


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D13957.38047.patch
Type: text/x-patch
Size: 3316 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151021/86e46b31/attachment.bin>