Branch: refs/heads/master
Home: https://github.com/dyninst/dyninst
Commit: f8ae92029312a1c9350ec81df9ee10cdba9dcdc7
https://github.com/dyninst/dyninst/commit/f8ae92029312a1c9350ec81df9ee10cdba9dcdc7
Author: bbiiggppiigg <bbiiggppi...@gmail.com>
Date: 2021-04-04 (Sun, 04 Apr 2021)
Changed paths:
M common/h/amdgpu_op_table.h
M common/h/amdgpu_vega_sys_regs.h
A dataflowAPI/rose/SgAsmAmdgpuVegaInstruction.h
M dataflowAPI/rose/conversions.h
M dataflowAPI/rose/semantics/BaseSemantics2.C
A dataflowAPI/rose/semantics/DispatcherAmdgpuVega.C
A dataflowAPI/rose/semantics/DispatcherAmdgpuVega.h
M dataflowAPI/rose/semantics/Registers.C
M dataflowAPI/rose/semantics/Registers.h
M dataflowAPI/rose/semantics/SymEvalSemantics.C
M dataflowAPI/rose/semantics/SymEvalSemantics.h
M dataflowAPI/src/AbslocInterface.C
M dataflowAPI/src/ExpressionConversionVisitor.C
M dataflowAPI/src/RoseImpl.C
M dataflowAPI/src/RoseInsnFactory.C
M dataflowAPI/src/RoseInsnFactory.h
M dataflowAPI/src/SymEval.C
M dataflowAPI/src/SymbolicExpansion.C
M dataflowAPI/src/SymbolicExpansion.h
M dataflowAPI/src/convertOpcodes.C
M external/rose/amdgpuInstructionEnum.h
M external/rose/rose-compat.h
M instructionAPI/src/Instruction.C
M instructionAPI/src/InstructionDecoder-amdgpu-vega.C
M instructionAPI/src/amdgpu_opcode_tables.C
M parseAPI/CMakeLists.txt
M parseAPI/src/IA_IAPI.C
M parseAPI/src/IndirectAnalyzer.C
M parseAPI/src/JumpTableFormatPred.C
M parseAPI/src/JumpTableFormatPred.h
M parseAPI/src/Parser.C
M parseAPI/src/SymbolicExpression.C
Log Message:
-----------
Add Initial support for Analyzing Indirect Control Flow on Amdgpu Vega (#979)
* this patch should fix most of the compile warnings
* add dependency to scc
* add support for negate and equal to zero op, and chopoff overflowed bits in
constant_ast
* fix readRegister to use the actual reg size instead of fixed 64
* add support for scc
* fix the instruction semantics to include scc in their computation
* remove left over x86 cases in RoseInsnAmdgpuVegaFactory::massageOperands
* remove left over arm registers and function handles
* revert the use of printf back to parsing_printf
* revert back to use parsing printf
* comment out debug message
* fix the problem of zeroing out x86_64 constant asts with zero width
* patch noly 32 bit positive integer
* should fix the logic in simplifying constant ast
* add dummy entries to avoid buffer overflow
* fix ast simplification logic on ppc64
* fix a bug in extractOP that might genereate wrong value for constant of size
64
_______________________________________________
Dyninst-api mailing list
Dyninst-api@cs.wisc.edu
https://lists.cs.wisc.edu/mailman/listinfo/dyninst-api
