Skip to content

Handle integer overflow when parsing maps. #391

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
byroot merged 1 commit into from
Jun 10, 2026
+65 −20
Merged

Handle integer overflow when parsing maps. #391

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .github/workflows/ci.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ jobs:
fail-fast: false
matrix:
os: [ubuntu]
ruby: ['jruby-9.3', 'jruby-9.4', 'truffleruby']
ruby: ['jruby-9.4', 'truffleruby']
runs-on: ${{ matrix.os }}-latest
steps:
- uses: actions/checkout@v4
Expand Down
2 changes: 2 additions & 0 deletions ChangeLog
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
* Fix an integer overflow when parsing maps.

2026-06-09 1.8.2

* Fix `Buffer#clear` to properly reset memory chunks before adding them back to the pool.
Expand Down
22 changes: 15 additions & 7 deletions ext/java/org/msgpack/jruby/Decoder.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,6 +85,14 @@ public Decoder(Ruby runtime, Unpacker unpacker, byte[] bytes, int offset, int le
feed(bytes, offset, length);
}

private int getUnsignedInt() {
int size = buffer.getInt();
if (size < 0) {
throw runtime.newRaiseException(underflowErrorClass, "Size too large (limited to 2**31 for the Java version)");
}
return size;
}

public void feed(byte[] bytes) {
feed(bytes, 0, bytes.length);
}
Expand Down Expand Up @@ -200,7 +208,7 @@ public IRubyObject read_array_header() {
} else if (b == ARY16) {
return runtime.newFixnum(buffer.getShort() & 0xffff);
} else if (b == ARY32) {
return runtime.newFixnum(buffer.getInt());
return runtime.newFixnum(getUnsignedInt());
}
throw runtime.newRaiseException(unexpectedTypeErrorClass, "unexpected type");
} catch (RaiseException re) {
Expand All @@ -221,7 +229,7 @@ public IRubyObject read_map_header() {
} else if (b == MAP16) {
return runtime.newFixnum(buffer.getShort() & 0xffff);
} else if (b == MAP32) {
return runtime.newFixnum(buffer.getInt());
return runtime.newFixnum(getUnsignedInt());
}
throw runtime.newRaiseException(unexpectedTypeErrorClass, "unexpected type");
} catch (RaiseException re) {
Expand Down Expand Up @@ -258,10 +266,10 @@ private IRubyObject consumeNext() {
case TRUE: return runtime.getTrue();
case BIN8: return consumeString(buffer.get() & 0xff, binaryEncoding);
case BIN16: return consumeString(buffer.getShort() & 0xffff, binaryEncoding);
case BIN32: return consumeString(buffer.getInt(), binaryEncoding);
case BIN32: return consumeString(getUnsignedInt(), binaryEncoding);
case VAREXT8: return consumeExtension(buffer.get() & 0xff);
case VAREXT16: return consumeExtension(buffer.getShort() & 0xffff);
case VAREXT32: return consumeExtension(buffer.getInt());
case VAREXT32: return consumeExtension(getUnsignedInt());
case FLOAT32: return runtime.newFloat(buffer.getFloat());
case FLOAT64: return runtime.newFloat(buffer.getDouble());
case UINT8: return runtime.newFixnum(buffer.get() & 0xffL);
Expand All @@ -283,11 +291,11 @@ private IRubyObject consumeNext() {
case FIXEXT16: return consumeExtension(16);
case STR8: return consumeString(buffer.get() & 0xff, utf8Encoding);
case STR16: return consumeString(buffer.getShort() & 0xffff, utf8Encoding);
case STR32: return consumeString(buffer.getInt(), utf8Encoding);
case STR32: return consumeString(getUnsignedInt(), utf8Encoding);
case ARY16: return consumeArray(buffer.getShort() & 0xffff);
case ARY32: return consumeArray(buffer.getInt());
case ARY32: return consumeArray(getUnsignedInt());
case MAP16: return consumeHash(buffer.getShort() & 0xffff);
case MAP32: return consumeHash(buffer.getInt());
case MAP32: return consumeHash(getUnsignedInt());
default: break outer;
}
case 0xe:
Expand Down
24 changes: 12 additions & 12 deletions ext/msgpack/unpacker.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -456,13 +456,13 @@ static int read_primitive(msgpack_unpacker_t* uk)
return object_complete(uk, INT2NUM((int8_t)b));

SWITCH_RANGE(b, 0xa0, 0xbf) // FixRaw / fixstr
int count = b & 0x1f;
size_t count = b & 0x1f;
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_STRING);

SWITCH_RANGE(b, 0x90, 0x9f) // FixArray
int count = b & 0x0f;
size_t count = b & 0x0f;
if(count == 0) {
return object_complete(uk, rb_ary_new());
}
Expand Down Expand Up @@ -638,7 +638,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xd9: // raw 8 / str 8
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 1);
uint8_t count = cb.u8;
size_t count = cb.u8;
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_STRING);
Expand All @@ -647,7 +647,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xda: // raw 16 / str 16
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 2);
uint16_t count = _msgpack_be16(cb.u16);
size_t count = _msgpack_be16(cb.u16);
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_STRING);
Expand All @@ -656,7 +656,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xdb: // raw 32 / str 32
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 4);
uint32_t count = _msgpack_be32(cb.u32);
size_t count = _msgpack_be32(cb.u32);
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_STRING);
Expand All @@ -665,7 +665,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xc4: // bin 8
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 1);
uint8_t count = cb.u8;
size_t count = cb.u8;
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_BINARY);
Expand All @@ -674,7 +674,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xc5: // bin 16
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 2);
uint16_t count = _msgpack_be16(cb.u16);
size_t count = _msgpack_be16(cb.u16);
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_BINARY);
Expand All @@ -683,7 +683,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xc6: // bin 32
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 4);
uint32_t count = _msgpack_be32(cb.u32);
size_t count = _msgpack_be32(cb.u32);
/* read_raw_body_begin sets uk->reading_raw */
uk->reading_raw_remaining = count;
return read_raw_body_begin(uk, RAW_TYPE_BINARY);
Expand All @@ -692,7 +692,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xdc: // array 16
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 2);
uint16_t count = _msgpack_be16(cb.u16);
size_t count = _msgpack_be16(cb.u16);
if(count == 0) {
return object_complete(uk, rb_ary_new());
}
Expand All @@ -702,7 +702,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xdd: // array 32
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 4);
uint32_t count = _msgpack_be32(cb.u32);
size_t count = _msgpack_be32(cb.u32);
if(count == 0) {
return object_complete(uk, rb_ary_new());
}
Expand All @@ -712,7 +712,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xde: // map 16
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 2);
uint16_t count = _msgpack_be16(cb.u16);
size_t count = _msgpack_be16(cb.u16);
if(count == 0) {
return object_complete(uk, rb_hash_new());
}
Expand All @@ -722,7 +722,7 @@ static int read_primitive(msgpack_unpacker_t* uk)
case 0xdf: // map 32
{
READ_CAST_BLOCK_OR_RETURN_EOF(cb, uk, 4);
uint32_t count = _msgpack_be32(cb.u32);
size_t count = _msgpack_be32(cb.u32);
if(count == 0) {
return object_complete(uk, rb_hash_new());
}
Expand Down
35 changes: 35 additions & 0 deletions spec/unpack_spec.rb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,10 @@
end

describe MessagePack do
let(:eof_error) do
IS_JRUBY ? MessagePack::UnderflowError : EOFError
end

it 'MessagePack.unpack symbolize_keys' do
symbolized_hash = {:a => 'b', :c => 'd'}
MessagePack.load(MessagePack.pack(symbolized_hash), :symbolize_keys => true).should == symbolized_hash
Expand Down Expand Up @@ -54,4 +58,35 @@
MessagePack.unpack([0xc6, 0x00, 0x00, 0x00, 0x01].pack('C*') + 'a').should == "a"
MessagePack.unpack([0xc6, 0x00, 0x00, 0x00, 0x02].pack('C*') + 'aa').should == "aa"
end

it "msgpack fixmap type" do
MessagePack.unpack([0x81, 0xa1, 0x61, 0x01].pack('C*')).should == {"a" => 1}
expect {
MessagePack.unpack([0x82, 0xa1, 0x61, 0x01].pack('C*'))
}.to raise_error(eof_error)
end

it "msgpack map 16 type" do
MessagePack.unpack([0xde, 0x00, 0x01, 0xa1, 0x61, 0x1].pack('C*')).should == {"a" => 1}

expect {
MessagePack.unpack([0xde, 0x00, 0x02, 0xa1, 0x61, 0x1].pack('C*'))
}.to raise_error(eof_error)

expect {
MessagePack.unpack([0xde, 0x80, 0x01, 0xa1, 0x61, 0x1].pack('C*'))
}.to raise_error(eof_error)
end

it "msgpack map 32 type" do
MessagePack.unpack([0xdf, 0x00, 0x00, 0x00, 0x01, 0xa1, 0x61, 0x1].pack('C*')).should == {"a" => 1}

expect {
MessagePack.unpack([0xdf, 0x00, 0x00, 0x00, 0x02, 0xa1, 0x61, 0x1].pack('C*'))
}.to raise_error(eof_error)

expect {
p MessagePack.unpack([0xdf, 0x80, 0x00, 0x00, 0x01, 0xa1, 0x61, 0x1].pack('C*'))
}.to raise_error(eof_error)
end
end
Loading