reven2.trace.Transition

Information

Information

Inverse operations

The transition switches between user and kernel land. Examples:

• a syscall transition => the related sysret transition
• a sysret transition => the related syscall transition
• a exception transition => the related iretq transition
• a iretq transition => the related exception transition

The transition does memory accesses:

• case 1: a unique access. The access is selected.
• case 2: multiple write accesses. The first one is selected.
• case 3: multiple read accesses. The first one is selected.
• case 4: multiple read and write accesses. The first write access is selected. This enable to get the matching ret transition on an indirect call transition e.g. call [rax + 10].

If the selected access is a write then the next read access on the same memory range is searched for.

If the selected access is a read then the previous write access on the same memory range is searched for.

Example find_inverse on:

• a call transition => the related ret transition.
• a ret transition => the related call transition.
• a push transition => the related pop or mov transition.
• a pop transition => the related push transition.
• a store transition => the related load transition.
• a load transition => the related store transition.

Note

Due to the fact that find_inverse matches with memory accesses to find the inverse instruction, in some special cases, such as ROP chain, the inverse of e.g. a ret will not be a call instruction, but could be a mov to the memory, for example.

Dependencies

This method requires that the Reven2 server have the Memory history enabled.

Usage

It can be combined with other features like backtrace to obtain interesting results.

Example

For example, to jump to the end of the current function:

>>> import reven2
>>> reven_server = reven2.RevenServer('localhost', 13370)
>>> current_transition = reven_server.trace.transition(10000000)
>>> ret_transition = current_transition.find_inverse()

Information

Information

Warning

The link only works using classic Jupyter Notebook. It does not work with the JupyterLab used by esReven.

Information

Examples

>>> # Getting all accesses as a list at transition 42 (can be long if there are a lots of accesses):
>>> list(trace.transition(42).memory_accesses())
[MemoryAccess(transition=Transition(id=42), physical_address=PhysicalAddress(offset=0x7fc03eb8), size=8,
    operation=MemoryAccessOperation.Write, virtual_address=LinearAddress(offset=0xffff88007fc03eb8))]

>>> # Getting the first memory access at transition 14
>>> next(trace.transition(14).memory_accesses())
MemoryAccess(transition=Transition(id=14), physical_address=PhysicalAddress(offset=0x1f270a2), size=1,
    operation=MemoryAccessOperation.Read, virtual_address=LinearAddress(offset=0xffffffff81f270a2))

>>> # Getting all addresses that are read at transition 0.
>>> addresses = set()
>>> for access in trace.transition(0).memory_accesses(operation=reven2.memhist.MemoryAccessOperation.Read):
...     physical_offset = access.physical_address.offset
...     for address in range(physical_offset, physical_offset + access.size):
...         addresses.add(address)
>>> for address in addresses:
...     print(reven2.address.PhysicalAddress(address))
phy:0x36f05080
phy:0x36f05081
phy:0x36f05082
phy:0x36f05083
phy:0x36f05084
phy:0x36f05085
phy:0x36f05086
phy:0x36f05087

Information

Information

Information