[
https://issues.apache.org/jira/browse/AVRO-4298?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ismaël Mejía updated AVRO-4298:
-------------------------------
Description:
A bytes or string value is encoded as a length prefix followed by that many
bytes of data, and an array or map block is encoded as an element count
followed by that many items. A malicious or truncated input can declare a very
large length or count while carrying little or no actual data, causing a large
allocation before the shortfall is noticed. When the source can report how many
bytes remain, reject a declared length (or a collection block count) that
exceeds the bytes actually available before allocating for it. Companion to
AVRO-4241 (Java).
AvroIOBinaryDecoder::bytesRemaining() backs the check; read() consults it
directly for a declared length above a threshold, while
AvroIODatumReader::readArray/readMap consult it via minBytesPerElement().
ensureCollectionAvailable enforces the per-block limits and the decoder's
skipArray/skipMap are bounded (element-aware). Collection size-cap rejections
raise the dedicated AvroIOCollectionSizeException; available-bytes and
malformed-input rejections raise AvroException.
Zero-byte elements (null, a zero-length fixed, or a record with only zero-byte
fields) consume no input, so the available-bytes check cannot bound their
count: a tiny payload such as {"type":"array","items":"null"} declaring a block
count of 200,000,000 would otherwise drive an unbounded allocation. In addition
to the available-bytes check this therefore caps the cumulative count of
zero-byte elements (default 10,000,000), applies a structural cap
(Integer.MAX_VALUE - 8) to every non-zero-byte-element collection (which also
covers collections read from a source that cannot report the bytes remaining),
and bounds the array/map skip paths. When set, the AVRO_MAX_COLLECTION_ITEMS
environment variable caps both limits. This supersedes the separate
collection-limit sub-task.
was:
A bytes or string value is encoded as a length prefix followed by that many
bytes of data, and an array or map block is encoded as an element count
followed by that many items. A malicious or truncated input can declare a very
large length or count while carrying little or no actual data, causing a large
allocation before the shortfall is noticed. When the source can report how many
bytes remain, reject a declared length (or a collection block count) that
exceeds the bytes actually available before allocating for it. Companion to
AVRO-4241 (Java).
AvroIOBinaryDecoder::bytesRemaining() backs the check; read() and
AvroIODatumReader::readArray/readMap consult it via minBytesPerElement().
ensureCollectionAvailable enforces the per-block limits and the decoder's
skipArray/skipMap are bounded (element-aware). Rejections raise the dedicated
AvroIOCollectionSizeException.
Zero-byte elements (null, a zero-length fixed, or a record with only zero-byte
fields) consume no input, so the available-bytes check cannot bound their
count: a tiny payload such as {"type":"array","items":"null"} declaring a block
count of 200,000,000 would otherwise drive an unbounded allocation. In addition
to the available-bytes check this therefore caps the cumulative count of
zero-byte elements (default 10,000,000), applies a structural cap to every
collection (Integer.MAX_VALUE - 8) for readers that cannot report bytes
remaining, and bounds the array/map skip paths. When set, the
AVRO_MAX_COLLECTION_ITEMS environment variable caps both limits. This
supersedes the separate collection-limit sub-task.
> [php] Bound allocation when decoding length-prefixed values and collections
> ---------------------------------------------------------------------------
>
> Key: AVRO-4298
> URL: https://issues.apache.org/jira/browse/AVRO-4298
> Project: Apache Avro
> Issue Type: Sub-task
> Components: php
> Affects Versions: 1.11.5, 1.12.1
> Reporter: Ismaël Mejía
> Assignee: Ismaël Mejía
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.13.0, 1.11.6, 1.12.2
>
> Time Spent: 4h 20m
> Remaining Estimate: 0h
>
> A bytes or string value is encoded as a length prefix followed by that many
> bytes of data, and an array or map block is encoded as an element count
> followed by that many items. A malicious or truncated input can declare a
> very large length or count while carrying little or no actual data, causing a
> large allocation before the shortfall is noticed. When the source can report
> how many bytes remain, reject a declared length (or a collection block count)
> that exceeds the bytes actually available before allocating for it. Companion
> to AVRO-4241 (Java).
> AvroIOBinaryDecoder::bytesRemaining() backs the check; read() consults it
> directly for a declared length above a threshold, while
> AvroIODatumReader::readArray/readMap consult it via minBytesPerElement().
> ensureCollectionAvailable enforces the per-block limits and the decoder's
> skipArray/skipMap are bounded (element-aware). Collection size-cap rejections
> raise the dedicated AvroIOCollectionSizeException; available-bytes and
> malformed-input rejections raise AvroException.
> Zero-byte elements (null, a zero-length fixed, or a record with only
> zero-byte fields) consume no input, so the available-bytes check cannot bound
> their count: a tiny payload such as {"type":"array","items":"null"} declaring
> a block count of 200,000,000 would otherwise drive an unbounded allocation.
> In addition to the available-bytes check this therefore caps the cumulative
> count of zero-byte elements (default 10,000,000), applies a structural cap
> (Integer.MAX_VALUE - 8) to every non-zero-byte-element collection (which also
> covers collections read from a source that cannot report the bytes
> remaining), and bounds the array/map skip paths. When set, the
> AVRO_MAX_COLLECTION_ITEMS environment variable caps both limits. This
> supersedes the separate collection-limit sub-task.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
