# XML Billion Laughs Entity Expansion via lxml.etree.XMLParser(resolve_entities=True)

Language: Python
Severity: High
CWE: CWE-611

## Source
3

## Flow
3-6-8

## Sink
8

## Vulnerable Code
```python
from lxml import etree
import boto3

def process_iot_device_manifest(manifest_xml, device_id):
    s3_client = boto3.client('s3')
    xml_parser = etree.XMLParser(resolve_entities=True, no_network=False)
    try:
        device_config = etree.fromstring(manifest_xml.encode(), parser=xml_parser)
        firmware_version = device_config.find('.//firmware').text
        telemetry_interval = device_config.find('.//telemetry_interval').text
        s3_client.put_object(Bucket='iot-device-configs', Key=f'devices/{device_id}/manifest.xml', Body=manifest_xml)
        return {'device_id': device_id, 'firmware': firmware_version, 'interval': telemetry_interval}
    except etree.XMLSyntaxError as err:
        return {'error': 'Invalid manifest format', 'device_id': device_id}
```

## Explanation

The function accepts untrusted XML input (manifest_xml) from IoT devices and parses it with resolve_entities=True, enabling XML External Entity (XXE) and Billion Laughs attacks. The parser expands entities before validation, allowing attackers to cause denial of service through exponential entity expansion or exfiltrate local files.

## Remediation

The fix disables entity resolution by setting resolve_entities=False, prevents network access with no_network=True, and disables DTD loading and validation with dtd_validation=False and load_dtd=False. This prevents both Billion Laughs (exponential entity expansion) and XXE attacks by ensuring no entities are resolved and no external resources are fetched during parsing.

## Secure Code
```python
from lxml import etree
import boto3

def process_iot_device_manifest(manifest_xml, device_id):
    s3_client = boto3.client('s3')
    xml_parser = etree.XMLParser(resolve_entities=False, no_network=True, dtd_validation=False, load_dtd=False)
    try:
        device_config = etree.fromstring(manifest_xml.encode(), parser=xml_parser)
        firmware_version = device_config.find('.//firmware').text
        telemetry_interval = device_config.find('.//telemetry_interval').text
        s3_client.put_object(Bucket='iot-device-configs', Key=f'devices/{device_id}/manifest.xml', Body=manifest_xml)
        return {'device_id': device_id, 'firmware': firmware_version, 'interval': telemetry_interval}
    except etree.XMLSyntaxError as err:
        return {'error': 'Invalid manifest format', 'device_id': device_id}
```
