.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "api/gallery/vm/ethereum/plot_03_reentrancy.py"
.. LINE NUMBERS ARE GIVEN BELOW.

.. only:: html

    .. note::
        :class: sphx-glr-download-link-note

        :ref:`Go to the end <sphx_glr_download_api_gallery_vm_ethereum_plot_03_reentrancy.py>`
        to download the full example code or to run this example in your browser via JupyterLite.

.. rst-class:: sphx-glr-example-title

.. _sphx_glr_api_gallery_vm_ethereum_plot_03_reentrancy.py:


The DAO and reentrancy: checks, effects, interactions (2016)
============================================================

The DAO was an Ethereum fund holding about 14% of all ether. Its withdrawal
code sent ether *before* updating the caller's balance. Sending ether to a
contract runs the contract's code, so in June 2016 an attacker's contract
called back into the withdrawal from inside the payment, again and again,
while its balance still showed the original deposit. About 3.6 million
ether were drained, and Ethereum hard-forked to return them.

What to look for
----------------

With the vulnerable order, the attacker withdraws its deposit once per
nested call, until the bank is empty or the call-depth limit stops it. With
the *checks-effects-interactions* order, the re-entrant call finds a zero
balance and the attacker gets back exactly its deposit.

The history behind this experiment: :doc:`/history/vm_breakthroughs`.

.. GENERATED FROM PYTHON SOURCE LINES 24-26

Drain a bank holding 10,000 of other people's funds
---------------------------------------------------

.. GENERATED FROM PYTHON SOURCE LINES 26-40

.. code-block:: Python

    import matplotlib.pyplot as plt

    import blockchainkit as bk

    deposits = [10, 50, 100, 500, 1000]
    vulnerable = [bk.vm.drain_bank(10_000, d) for d in deposits]
    safe = [bk.vm.drain_bank(10_000, d, checks_effects_interactions=True) for d in deposits]
    for d, v, s in zip(deposits, vulnerable, safe, strict=True):
        print(f"deposit {d:5d}: vulnerable steals {v.stolen} in {v.calls} calls, safe {s.stolen}")
    assert all(s.stolen == 0 for s in safe)
    assert all(
        v.withdrawn == ((10_000 + d) // d) * d for d, v in zip(deposits, vulnerable, strict=True)
    )





.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    deposit    10: vulnerable steals 10000 in 1002 calls, safe 0
    deposit    50: vulnerable steals 10000 in 202 calls, safe 0
    deposit   100: vulnerable steals 10000 in 102 calls, safe 0
    deposit   500: vulnerable steals 10000 in 22 calls, safe 0
    deposit  1000: vulnerable steals 10000 in 12 calls, safe 0




.. GENERATED FROM PYTHON SOURCE LINES 41-44

The depth limit
---------------
Ethereum then limited calls to a depth of 1024; tiny deposits run into it.

.. GENERATED FROM PYTHON SOURCE LINES 44-63

.. code-block:: Python

    deep = bk.vm.drain_bank(10**9, 10)
    print(deep)
    assert deep.calls == 1024 and deep.withdrawn == 10_240

    fig, (left, right) = plt.subplots(1, 2, figsize=(10, 4))
    left.bar(
        ["pay, then update", "update, then pay"],
        [vulnerable[2].stolen, safe[2].stolen],
        color=["#dc2626", "#16a34a"],
    )
    left.set(ylabel="stolen", title="Deposit 100, bank 10,000")
    right.loglog(deposits, [v.calls for v in vulnerable], "o-", color="#dc2626")
    right.set(
        xlabel="attacker deposit",
        ylabel="nested withdraw calls",
        title="Smaller deposit, deeper recursion",
    )
    fig.tight_layout()




.. image-sg:: /api/gallery/vm/ethereum/images/sphx_glr_plot_03_reentrancy_001.png
   :alt: Deposit 100, bank 10,000, Smaller deposit, deeper recursion
   :srcset: /api/gallery/vm/ethereum/images/sphx_glr_plot_03_reentrancy_001.png
   :class: sphx-glr-single-img


.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    ReentrancyResult(withdrawn=10240, deposited=10, calls=1024, bank_balance=999989770)




.. GENERATED FROM PYTHON SOURCE LINES 64-69

Exercise
--------
A reentrancy *guard* sets a lock flag on entry and refuses to run while it
is set. Add a ``guarded`` option to a copy of ``drain_bank``'s loop. Why
do auditors still recommend checks-effects-interactions as well?


.. rst-class:: sphx-glr-timing

   **Total running time of the script:** (0 minutes 0.267 seconds)


.. _sphx_glr_download_api_gallery_vm_ethereum_plot_03_reentrancy.py:

.. only:: html

  .. container:: sphx-glr-footer sphx-glr-footer-example

    .. container:: lite-badge

      .. image:: images/jupyterlite_badge_logo.svg
        :target: ../../../../lite/lab/index.html?path=api/gallery/vm/ethereum/plot_03_reentrancy.ipynb
        :alt: Launch JupyterLite
        :width: 150 px

    .. container:: sphx-glr-download sphx-glr-download-jupyter

      :download:`Download Jupyter notebook: plot_03_reentrancy.ipynb <plot_03_reentrancy.ipynb>`

    .. container:: sphx-glr-download sphx-glr-download-python

      :download:`Download Python source code: plot_03_reentrancy.py <plot_03_reentrancy.py>`

    .. container:: sphx-glr-download sphx-glr-download-zip

      :download:`Download zipped: plot_03_reentrancy.zip <plot_03_reentrancy.zip>`


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_