[PATCH] D29134: AMDGPU: Set calling_convention bit in kernel_code_t

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
Herald added a reviewer: tstellarAMD.
Herald added subscribers: tony-tye, yaxunl, nhaehnle, wdng, kzhuravl.

According to the documentation this is supposed to be -1
if indirect calls are not supported.


https://reviews.llvm.org/D29134

Files:
  lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
  test/CodeGen/AMDGPU/hsa.ll


Index: test/CodeGen/AMDGPU/hsa.ll
===================================================================
--- test/CodeGen/AMDGPU/hsa.ll
+++ test/CodeGen/AMDGPU/hsa.ll
@@ -45,6 +45,8 @@
 ; HSA: .amd_kernel_code_t
 ; HSA: enable_sgpr_private_segment_buffer = 1
 ; HSA: enable_sgpr_kernarg_segment_ptr = 1
+; HSA: wavefront_size = 6
+; HSA: call_convention = -1
 ; HSA: .end_amd_kernel_code_t
 ; HSA: s_load_dwordx2 s[{{[0-9]+:[0-9]+}}], s[4:5], 0x0
 
Index: lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
===================================================================
--- lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
+++ lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
@@ -127,6 +127,11 @@
   Header.kernel_code_entry_byte_offset = sizeof(Header);
   // wavefront_size is specified as a power of 2: 2^6 = 64 threads.
   Header.wavefront_size = 6;
+
+  // If the code object does not support indirect functions, then the value must
+  // be 0xffffffff.
+  Header.call_convention = -1;
+
   // These alignment values are specified in powers of two, so alignment =
   // 2^n.  The minimum alignment is 2^4 = 16.
   Header.kernarg_segment_alignment = 4;


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D29134.85776.patch
Type: text/x-patch
Size: 1139 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170125/7478ca29/attachment.bin>