Running the ARM instruction emulation test on a big-endian system
would fail, since the code doesn't respect endianness properly.
In EmulateInstructionARM::TestEmulation, code assumes that an
instruction opcode read in from the test file is in target byte
order, but it was in fact read in in host byte order.
More difficult to fix, the EmulationStateARM structure models
the overlapping sregs and dregs by a union in _sd_regs. This
only works correctly if the host is a little-endian system.
I've removed the union in favor of a simple array containing
the 32 sregs, and changed any code accessing dregs to explicitly
use the correct two sregs overlaying that dreg in the proper
target order.
Also, the EmulationStateARM::ReadPseudoMemory and WritePseudoMemory
track memory as a map of uint32_t values in host byte order, and
implement 64-bit memory accessing by splitting them up into two
uint32_t ones. However, callers expect memory contents to be
provided in the form of a byte array (in target byte order).
This means the uint32_t contents need to be byte-swapped on
BE systems, and when splitting up a 64-bit access into two 32-bit
ones, byte order has to be respected.
Differential Revision: http://reviews.llvm.org/D18984
llvm-svn: 266314
Scalar::GetBytes provides a non-const access to the underlying bytes
of the scalar value, supposedly allowing for modification of those
bytes. However, even with the current implementation, this is not
really possible. For floating-point scalars, the pointer returned
by GetBytes refers to a temporary copy; modifications to that copy
will be simply ignored. For integer scalars, the pointer refers
to internal memory of the APInt implementation, which isn't
supposed to be directly modifyable; GetBytes simply casts aways
the const-ness of the pointer ...
With my upcoming patch to fix Scalar::GetBytes for big-endian
systems, this problem is going to get worse, since there we need
temporary copies even for some integer scalars. Therefore, this
patch makes Scalar::GetBytes const, fixing all those problems.
As a follow-on change, RegisterValues::GetBytes must be made const
as well. This in turn means that the way of initializing a
RegisterValue by doing a SetType followed by writing to GetBytes
no longer works. Instead, I've changed SetValueFromData to do
the equivalent of SetType itself, and then re-implemented
SetFromMemoryData to work on top of SetValueFromData.
There is still a need for RegisterValue::SetType, since some
platform-specific code uses it to reinterpret the contents of
an already filled RegisterValue. To make this usage work in
all cases (even changing from a type implemented via Scalar
to a type implemented as a byte buffer), SetType now simply
copies the old contents out, and then reloads the RegisterValue
from this data using the new type via SetValueFromData.
This in turn means that there is no remaining caller of
Scalar::SetType, so it can be removed.
The only other follow-on change was in MIPS EmulateInstruction
code, where some uses of RegisterValue::GetBytes could be made
const trivially.
Differential Revision: http://reviews.llvm.org/D18980
llvm-svn: 266310
in thumb mode into one method in ArchSpec, replace checks for
specific cores in the disassembler with calls to this. Also call
this from the arm instruction emulation code.
The determination of whether a given ArchSpec is thumb-only is still
a bit of a hack, but at least the hack is consolidated into a single
place. In my original version of this patch http://reviews.llvm.org/D13578
I was calling into llvm's feature arm feature tables to make this
determination, like
#include "llvm/Support/TargetRegistry.h"
#include "llvm/MC/MCSubtargetInfo.h"
#include "llvm/../../lib/Target/ARM/ARMGenRegisterInfo.inc"
#include "llvm/../../lib/Target/ARM/ARMFeatures.h"
[...]
std::string triple (GetTriple().getTriple());
const char *cpu = "";
const char *features_str = "";
const llvm::Target *curr_target = llvm::TargetRegistry::lookupTarget(triple.c_str(), Error);
std::unique_ptr<llvm::MCSubtargetInfo> subtarget_info_up (curr_target->createMCSubtargetInfo(triple.c_str(), cpu, features_str));
if (subtarget_info_up->getFeatureBits()[llvm::ARM::FeatureNoARM])
{
return true;
}
but those tables are post-llvm-build generated and linking against them
for all of our different build system methods was a big hiccup that I
haven't had time to revisit convincingly.
I'll keep that reviews.llvm.org patch around to remind myself that I
need to take another run at linking against the necessary tables
again in llvm.
<rdar://problem/23022803>
llvm-svn: 265377
The IT instruction can specify condition code for up to 4 consecutive
instruction and it is used quite often by clang in epilogues causing
an issue when trying to unwind from locations covered by the IT
instruction and for locatins inmediately after the IT instruction.
Changes made to fix it:
* Introduce the concept of conditional instruction block what is a list
of consecutive instructions with the same condition. We update the
unwind information during the conditional instruction block and when
we reach the end of it (first instruction with a differemt condition)
then we restore the unwind information we had before the condition.
* Fix a bug in the ARM instruction emulator where neither PC nor the
ITSTATE was advanced when we reached an instruction what we can't
decode.
After the change we have no regression on android-arm running the
regular test suit and TestStandardUnwind also passes when running it
with clang as the compiler (previously it failed on an IT instruction).
Differential revision: http://reviews.llvm.org/D16814
llvm-svn: 260368
The ARM instruction emulator had 2 bugs related to the handling of the
IT instruction causing an error in single stepping:
* We haven't initialized the IT mask from the CPSR so if the last
instruction of the IT block is a branch and the condition is false
then the emulator evaluated the branch what resulted in an incorrect
pc for the next instruction.
* The ITSTATE was advanced before the execution of each instruction. As
a result the emulator was using the condition of following instruction
in every case. The ITSTATE should be edvanced after the execution of
an instruction except after an IT instruction.
Differential revision: http://reviews.llvm.org/D16772
llvm-svn: 259509
SUMMARY:
This patch merges emulation of similar instructions into a single function (wherever possible) to remove code duplication.
Reviewers: clayborg
Subscribers: mohit.bhakkad, sagar, jaydeep, lldb-commits
Differential Revision: http://reviews.llvm.org/D16051
llvm-svn: 257442
SUMMARY:
This patch merges emulation of similar instructions into a single function (wherever possible) to remove code duplication.
Reviewers: clayborg
Subscribers: mohit.bhakkad, sagar, jaydeep, lldb-commits
Differential Revision: http://reviews.llvm.org/D15886
llvm-svn: 256915
The incorrect instruction emulation caused issues in the stack unwinding
code when strd was used to push 2 register to the stack with writeback.
llvm-svn: 256000
in places where we check for Triple::IOS. They're mostly the same as far
as lldb is conerned.
.
Also add a base cass implementation for Process::IsAlive - Greg added this
last year but it didn't get upstreamed.
llvm-svn: 252227
It is a new function added to the llvm::Triple class to simplify the
checking if we are targeting android to clean up the confusion between
android being an OS or an environment.
llvm-svn: 250286
Summary: This was deprecated and removed.
Reviewers: clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D13463
llvm-svn: 249452
SUMMARY:
This patch includes:
1. Emulation of prologue/epilogue and branch instructions for microMIPS.
2. Setting up alternate disassembler (to be used for microMIPS).
So there will be two disassembler instances, one for microMIPS and other for MIPS.
Appropriate disassembler will be used based on the address class of instruction address.
3. Some of the branch instructions does not have fixed sized delay slot, that means delay slot instruction can be of 2-byte or 4-byte.
For this "m_next_inst_size" has been introduced which stores the size of next instruction (i.e size of delay slot instruction in case of branch).
This can be used wherever the size of next instruction is required.
4. A minor change to use mips32 register names instead of mips64 names.
Reviewers: clayborg, tberghammer
Subscribers: mohit.bhakkad, sagar, jaydeep, nitesh.jain, lldb-commits
Differential Revision: http://reviews.llvm.org/D13282
llvm-svn: 249381
"gcc" register numbers are now correctly referred to as "ehframe"
register numbers. In almost all cases, ehframe and dwarf register
numbers are identical (the one exception is i386 darwin where ehframe
regnums were incorrect).
The old "gdb" register numbers, which I incorrectly thought were
stabs register numbers, are now referred to as "Process Plugin"
register numbers. This is the register numbering scheme that the
remote process controller stub (lldb-server, gdbserver, core file
support, kdp server, remote jtag devices, etc) uses to refer to the
registers. The process plugin register numbers may not be contiguous
- there are remote jtag devices that have gaps in their register
numbering schemes.
I removed all of the enums for "gdb" register numbers that we had
in lldb - these were meaningless - and I put LLDB_INVALID_REGNUM
in all of the register tables for the Process Plugin regnum slot.
This change is almost entirely mechnical; the one actual change in
here is to ProcessGDBRemote.cpp's ParseRegisters() which parses the
qXfer:features:read:target.xml response. As it parses register
definitions from the xml, it will assign sequential numbers as the
eRegisterKindLLDB numbers (the lldb register numberings must be
sequential, without any gaps) and if the xml file specifies register
numbers, those will be used as the eRegisterKindProcessPlugin
register numbers (and those may have gaps). A J-Link jtag device's
target.xml does contain a gap in register numbers, and it only
specifies the register numbers for the registers after that gap.
The device supports many different ARM boards and probably selects
different part of its register file as appropriate.
http://reviews.llvm.org/D12791
<rdar://problem/22623262>
llvm-svn: 247741
Summary:
This doesn't exist in other LLVM projects any longer and doesn't
do anything.
Reviewers: chaoren, labath
Subscribers: emaste, tberghammer, lldb-commits, danalbert
Differential Revision: http://reviews.llvm.org/D12586
llvm-svn: 246749
Change the way EmulateInstruction::eContextPopRegisterOffStack handled
in UnwindAssemblyInstEmulation::WriteRegister to accomodate for
additional cases when eContextPopRegisterOffStack (pop PC/FLAGS).
llvm-svn: 245690
On ARM there is no difference petween a pop and a load instruction so
a register can be loaded multiple times during the function. Add check
to threat the load as a restore only if it do the restore from the
same location where the register was saved.
Differential revision: http://reviews.llvm.org/D11947
llvm-svn: 245546
This patch :
- Fixes offsets of all register sets for Mips.
- Adds MSA register set and FRE=1 mode support for FP register set.
- Separates lldb register numbers and register infos of freebsd/mips64 from linux/mips64.
- Re-orders the register numbers of all kinds for mips to be consistent with freebsd order of register numbers.
Reviewers: jaydeep, clayborg, jasonmolenda, ovyalov, emaste
Subscribers: tberghammer, ovyalov, emaste, mohit.bhakkad, nitesh.jain, bhushan
Differential: http://reviews.llvm.org/D10919
llvm-svn: 245217
SUMMARY:
Last 3bits of the watchpoint address are masked by the kernel. For example, n is
at 0x120010d00 and m is 0x120010d04. When a watchpoint is set at m, then watch
exception is generated even when n is read/written. To handle this case, instruction
at PC is emulated to find the base address of the load/store instruction. This address
is then appended to the description of the stop-info packet. Client then reads this
information to check whether the user has set a watchpoint on this address.
Reviewers: jingham, clayborg
Subscribers: nitesh.jain, mohit.bhakkad, sagar, bhushan and lldb-commits
Differential Revision: http://reviews.llvm.org/D11672
llvm-svn: 244864
This change :
- Fixes offsets of all register sets for Mips.
- Adds MSA register set and FRE=1 mode support for FP register set.
- Separates lldb register numbers and register infos of freebsd/mips64 from linux/mips64.
- Re-orders the register numbers of all kinds for mips to be consistent with freebsd order of register numbers.
- Eliminates ENABLE_128_BIT_SUPPORT and union ValueData from Scalar.cpp and uses llvm::APInt and llvm::APFloat for all integer and floating point types.
Reviewers : emaste, jaydeep, clayborg
Subscribers : emaste, mohit.bhakkad, nitesh.jain, bhushan
Differential : http://reviews.llvm.org/D10919
llvm-svn: 244308
Don't chane the CFI information when a conditional instruction
is emulated (eg.: popeq {r0, pc}) because the CFI for the next
instruction should be the same as the CFI for the current instruction.
Differential revision: http://reviews.llvm.org/D11258
llvm-svn: 242519
SUMMARY:
The patch detects MIPS application specific extensions (ASE) like micromips by reading
ELF header.e_flags and SHT_MIPS_ABIFLAGS section. MIPS triple does not contain ASE
information like micromips, mips16, DSP, MSA etc. These can be read from header.e_flags
or SHT_MIPS_ABIFLAGS section.
Reviewers: clayborg
Subscribers: mohit.bhakkad, sagar, lldb-commits
Differential Revision: http://reviews.llvm.org/D11133
llvm-svn: 242381
Summary: This aligns the library names used by the Makefile build to be the same as those create by the CMake build to make switching between the two easier. The only major difficulty was lldbHost which was one library in the CMake system and several in the Makefile system. Most of the other changes are trivial renames.
Reviewers: labath
Subscribers: emaste, tberghammer, lldb-commits
Differential Revision: http://reviews.llvm.org/D11154
llvm-svn: 242196
The emulation of the branches are required by the new stack
unwinding logic to reinstantiate the prologue at the right place.
Differential revision: http://reviews.llvm.org/D10702
llvm-svn: 240769
* Add and fix the emulation of several instruction.
* Disable frame pointer usage on Android.
* Specify return address register for the unwind plan instead of explict
tracking the value of RA.
* Replace prologue detection heuristics (unreliable in several cases)
with a logic to follow the branch instructions and restore the CFI
value based on them. The target address for a branch should have the
same CFI as the source address (if they are in the same function).
* Handle symbols in ELF files where the symbol size is not specified
with calcualting their size based on the next symbol (already done
in MachO files).
* Fix architecture in FuncUnwinders with filling up the inforamtion
missing from the object file with the architecture of the target.
* Add code to read register wehn the value is set to "IsSame" as it
meanse the value of a register in the parent frame is the same as the
value in the current frame.
Differential revision: http://reviews.llvm.org/D10447
llvm-svn: 240533
SUMMARY:
This patch implements
1. Emulation of MIPS32 branch instructions
2. Enable single-stepping for MIPS32 instructions
3. Correction in emulation of MIPS64 branch instructions with delay slot
4. Adjust breakpoint address when breakpoint is hit in a forbidden slot of compact branch instruction
Reviewers: clayborg
Subscribers: mohit.bhakkad, sagar, bhushan, lldb-commits, emaste, nitesh.jain
Differential Revision: http://reviews.llvm.org/D10596
llvm-svn: 240373
Patch by Jaydeep Patil
SUMMARY:
1. Added emulation of MIPS64 floating-point branch instructions
2. Updated GetRegisterInfo to recognize floating-point registers
3. Provided CPU information while creating createMCSubtargetInfo in disassembler
4. Bug fix in emulation of JIC and JIALC
5. Correct identification of breakpoint when set in a delay slot of a branch instruction
Reviewers: clayborg
Subscribers: bhushan, mohit.bhakkad, sagar, nitesh.jain, lldb-commits.
Differential Revision: http://reviews.llvm.org/D10355
llvm-svn: 239996