VTIL-Core/VTIL-Architecture/routine/basic_block.cpp
2020-08-23 23:41:29 +02:00

406 lines
11 KiB
C++

// Copyright (c) 2020 Can Boluk and contributors of the VTIL Project
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
//
// 1. Redistributions of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
// 2. Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
// 3. Neither the name of VTIL Project nor the names of its contributors
// may be used to endorse or promote products derived from this software
// without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
// POSSIBILITY OF SUCH DAMAGE.
//
#include "basic_block.hpp"
#include <vtil/amd64>
#include <vtil/arm64>
namespace vtil
{
// Creates a new block bound to a new routine with the given parameters.
//
basic_block* basic_block::begin( vip_t entry_vip, architecture_identifier arch_id )
{
// Caller must provide a valid virtual instruction pointer.
//
fassert( entry_vip != invalid_vip );
// Create a routine and invoke create block.
//
routine* rtn = new routine{ arch_id };
return rtn->create_block( entry_vip ).first;
}
// Creates a new block connected to this block at the given vip, if already explored returns nullptr,
// should still be called if the caller knowns it is explored since this function creates the linkage.
//
basic_block* basic_block::fork( vip_t entry_vip )
{
// Block cannot be forked before a branching instruction is hit.
//
fassert( is_complete() );
// Caller must provide a valid virtual instruction pointer.
//
fassert( entry_vip != invalid_vip );
// Invoke create block.
//
auto [blk, inserted] = owner->create_block( entry_vip, this );
return inserted ? blk : nullptr;
}
// Queues a stack shift.
//
basic_block* basic_block::shift_sp( int64_t offset, bool merge_instance, const const_iterator& it_const )
{
// Drop const qualifier of the iterator, since we are in a non-const
// qualified member function, this qualifier is unnecessary.
//
iterator it = acquire( it_const );
// If requested, shift the stack index first.
//
if ( merge_instance )
{
// Assert instruction at iterator indeed resets stack pointer.
//
fassert( !it.is_end() && it->sp_reset );
// Decrement stack index for each instruction afterwards.
//
for ( auto i = std::next( it ); !i.is_end(); i++ )
( +i )->sp_index--;
sp_index--;
// Remove the reset flag and merge the offsets.
//
( +it )->sp_reset = false;
offset += it->sp_offset;
( +it )->sp_offset = 0;
}
// If an iterator is provided, shift the stack pointer
// for every instruction that precedes it as well.
//
uint32_t shifted_spi = it.is_end() ? -1 : it->sp_index;
for ( ;!it.is_end() && it->sp_index == shifted_spi; it++ )
{
// Shift the stack offset accordingly.
//
( +it )->sp_offset += offset;
// If memory operation:
//
if ( it->base->accesses_memory() )
{
// If base is stack pointer, add the offset.
//
auto [base, off] = ( +it )->memory_location();
if ( base.is_stack_pointer() )
off += offset;
}
}
// If we've reached the end, shift the final block offset as well.
//
if( it.is_end() ) sp_offset += offset;
return this;
}
// Emits an entire instruction using series of VEMITs.
//
basic_block* basic_block::vemits( const std::string& assembly )
{
std::vector<uint8_t> bytes;
switch ( owner->arch_id )
{
case architecture_amd64: bytes = amd64::assemble( assembly ); break;
case architecture_arm64: bytes = arm64::assemble( assembly ); break;
default: unreachable();
}
fassert( !bytes.empty() );
for ( uint8_t byte : bytes )
vemit( byte );
return this;
}
// Pushes an operand up the stack queueing the
// shift in stack pointer.
//
basic_block* basic_block::push( const operand& op )
{
// Handle SP specially since we change the stack pointer
// before the instruction begins.
//
if ( op.is_register() && op.reg().is_stack_pointer() )
{
auto t0 = tmp( 64 );
return mov( t0, op )->push( t0 );
}
// If operand size is not aligned:
//
if ( size_t misalignment = op.size() % VTIL_ARCH_POPPUSH_ENFORCED_STACK_ALIGN )
{
// Adjust for misalignment and zero the padding.
//
int64_t padding_size = VTIL_ARCH_POPPUSH_ENFORCED_STACK_ALIGN - misalignment;
shift_sp( -padding_size );
str( REG_SP, sp_offset, operand( 0, math::narrow_cast<bitcnt_t>( padding_size * 8 ) ) );
}
// Shift and write the operand.
//
shift_sp( -int64_t( op.size() ) );
str( REG_SP, sp_offset, op );
return this;
}
// Pops an operand from the stack queueing the
// shift in stack pointer.
//
basic_block* basic_block::pop( const operand& op )
{
// Save the pre-shift offset.
//
int64_t offset = sp_offset;
// If operand size is not aligned:
//
if ( size_t misalignment = op.size() % VTIL_ARCH_POPPUSH_ENFORCED_STACK_ALIGN )
{
// Adjust for misalignment.
//
shift_sp( VTIL_ARCH_POPPUSH_ENFORCED_STACK_ALIGN - misalignment );
}
// Shift and read to the operand.
//
shift_sp( op.size() );
ldd( op, REG_SP, offset );
return this;
}
// Instruction deletion.
//
il_iterator basic_block::erase( const const_iterator& pos )
{
// Signal modification.
//
signal_modification();
// If no previous entry, head and possibly also tail:
//
list_entry* entry = pos.entry;
if ( !entry->prev )
{
// Set head, if valid fix prev link, otherwise fix tail.
//
if ( head = entry->next )
head->prev = nullptr;
else
tail = nullptr;
}
// If there is previous, but no next, tail:
//
else if ( !entry->next )
{
// Set new tail and fix next link.
//
tail = entry->prev;
tail->next = nullptr;
}
// Else generic entry, fix links:
//
else
{
entry->prev->next = entry->next;
entry->next->prev = entry->prev;
}
// Delete the entry and return next.
//
iterator npos = { this, entry->next };
destruct_instruction( entry );
instruction_count--;
return npos;
}
basic_block* basic_block::clear()
{
// Destruct every entry.
//
for ( auto it = head; it; )
{
auto next = it->next;
destruct_instruction( it );
it = next;
}
// Reset the state saved and return self.
//
head = nullptr;
tail = nullptr;
signal_modification();
instruction_count = 0;
return this;
}
instruction basic_block::pop_front()
{
// Save instruction at head and erase it.
//
dassert( head );
instruction result = std::move( head->value );
erase( { this, head } );
return result;
}
instruction basic_block::pop_back()
{
// Save instruction at tail and erase it.
//
dassert( tail );
instruction result = std::move( tail->value );
erase( { this, tail } );
return result;
}
// Internally invoked by emplace to insert a new linked list entry to the instruction stream.
//
il_iterator basic_block::insert_final( const const_iterator& pos, list_entry* new_entry, bool process )
{
// Signal modification.
//
signal_modification();
// If marked to be processed:
//
if( process )
{
// Validate instruction.
//
auto& ins = new_entry->value;
ins.is_valid( true );
// Validate registers are of matching architecture.
//
for ( auto& op : ins.operands )
{
if ( op.is_register() && op.reg().is_physical() && !op.reg().is_special() )
fassert( op.reg().architecture == owner->arch_id );
}
// If label stack is not empty and instruction has an invalid vip, use the last label pushed.
//
if ( !label_stack.empty() && ins.vip == invalid_vip )
ins.vip = label_stack.back();
// Instructions cannot be appended after a branching instruction was hit.
//
auto& it = acquire( pos );
if ( it.is_end() && !it.is_begin() )
fassert( !std::prev( it )->base->is_branching() );
// If inserting at end, inherit stack properties from the container.
//
if ( it.is_end() )
{
ins.sp_offset = sp_offset;
ins.sp_index = sp_index;
}
// If inserting at the beginning, assume clean stack state.
//
else if ( it.is_begin() )
{
ins.sp_offset = 0;
ins.sp_index = 0;
}
// If inserting in the middle of the stream:
//
else
{
// If previous instruction resets stack, use clean state of next index.
//
auto prev = std::prev( it );
if ( prev->sp_reset )
{
ins.sp_index = prev->sp_index + 1;
ins.sp_offset = 0;
}
// Otherwise inherit the state as is.
//
else
{
ins.sp_index = prev->sp_index;
ins.sp_offset = prev->sp_offset;
}
}
// If instruction writes to SP, reset the queued stack pointer.
//
for ( auto [op, type] : ins.enum_operands() )
{
if ( type >= operand_type::write && op.reg().is_stack_pointer() )
{
shift_sp( -ins.sp_offset, false, it );
for ( auto it2 = it; !it2.is_end(); it2++ )
( +it2 )->sp_index++;
sp_index++;
ins.sp_reset = true;
break;
}
}
}
// If iterator has a valid entry:
//
if ( pos.entry )
{
// Link before it.
//
new_entry->prev = pos.entry->prev;
new_entry->next = pos.entry;
pos.entry->prev = new_entry;
// Set head/tail if first entry, else fix links.
//
if ( new_entry->prev ) new_entry->prev->next = new_entry;
else head = new_entry;
}
// Otherwise, only entry or end.
//
else
{
// Link at the end, set tail.
//
new_entry->prev = tail;
new_entry->next = nullptr;
tail = new_entry;
// Set head if first entry, else fix links.
//
if ( new_entry->prev ) new_entry->prev->next = new_entry;
else head = new_entry;
}
// Increment entry count and return new iterator.
//
instruction_count++;
return { this, new_entry };
}
};