Files
scylladb/cql3/selection/selection.hh
Nadav Har'El c87d6407ed cql3: Replace SCYLLA_ASSERT by throwing_assert
In this patch we replace every single use of SCYLLA_ASSERT() in the cql3/
directory by throwing_assert().

The problem with SCYLLA_ASSERT() is that when it fails, it crashes Scylla.
This is almost always a bad idea (see #7871 discussing why), but it's even
riskier in front-end code like cql3/: In front-end code, there is a risk
that due to a bug in our code, a specific user request can cause Scylla
to crash. A malicious user can send this query to all nodes and crash
the entire cluster. When the user is not malicious, it causes a small
problem (a failing request) to become a much worse crash - and worse,
the user has no idea which request is causing this crash and the crash
will repeat if the same request is tried again.

All of this is solved by using the new throwing_assert(), which is the
same as SCYLLA_ASSERT() but throws an exception (using on_internal_error())
instead of crashing. The exception will prevent the code path with the
invalid assumption from continuing, but will result in only the current
user request being aborted, with a clear error message reporting the
internal server error due to an assertion failure.

I reviewed all the changes that I did in this patch to check that (to the
best of my understanding) none of the assertions in cql3/ involve the
sort of serious corruption that might require crashing the Scylla node
entirely.

throwing_assert() also improves logging of assertion failures compared
to the original SCYLLA_ASSERT() - SCYLLA_ASSERT() printed a message to
stderr which in many installations is lost, whereas throwing_assert()
uses Scylla's standard logger, and also includes a backtrace in the
log message.

Fixes #13970 (Exorcise assertions from CQL code paths)
Refs #7871 (Exorcise assertions from Scylla)

Signed-off-by: Nadav Har'El <nyh@scylladb.com>
2026-03-11 09:41:20 +02:00

405 lines
15 KiB
C++

/*
* Copyright (C) 2015-present ScyllaDB
*
* Modified by ScyllaDB
*/
/*
* SPDX-License-Identifier: (LicenseRef-ScyllaDB-Source-Available-1.0 and Apache-2.0)
*/
#pragma once
#include "utils/assert.hh"
#include "bytes.hh"
#include "schema/schema_fwd.hh"
#include "query/query-result-reader.hh"
#include "selector.hh"
#include "cql3/column_specification.hh"
#include "cql3/functions/function.hh"
#include "exceptions/exceptions.hh"
#include "unimplemented.hh"
#include <seastar/core/thread.hh>
namespace cql3 {
class result_set;
class metadata;
class query_options;
namespace restrictions {
class statement_restrictions;
}
namespace selection {
class raw_selector;
class result_set_builder;
class selectors {
public:
virtual ~selectors() {}
virtual bool requires_thread() const = 0;
virtual bool is_aggregate() const = 0;
/**
* Adds the current row of the specified <code>ResultSetBuilder</code>.
*
* @param rs the <code>ResultSetBuilder</code>
* @throws InvalidRequestException
*/
virtual void add_input_row(result_set_builder& rs) = 0;
virtual std::uint64_t get_input_row_count() const = 0;
virtual std::vector<managed_bytes_opt> get_output_row() = 0;
// When not aggregating, each input row becomes one output row.
virtual std::vector<managed_bytes_opt> transform_input_row(result_set_builder& rs) = 0;
virtual void reset() = 0;
};
class selection {
private:
schema_ptr _schema;
std::vector<const column_definition*> _columns;
::shared_ptr<metadata> _metadata;
const bool _collect_timestamps;
const bool _collect_TTLs;
const bool _contains_static_columns;
bool _is_trivial;
protected:
using trivial = bool_class<class trivial_tag>;
selection(schema_ptr schema,
std::vector<const column_definition*> columns,
std::vector<lw_shared_ptr<column_specification>> metadata_,
bool collect_timestamps,
bool collect_TTLs, trivial is_trivial = trivial::no);
virtual ~selection() {}
public:
// Overridden by SimpleSelection when appropriate.
virtual bool is_wildcard() const {
return false;
}
/**
* Checks if this selection contains static columns.
* @return <code>true</code> if this selection contains static columns, <code>false</code> otherwise;
*/
bool contains_static_columns() const {
return _contains_static_columns;
}
/**
* Checks if this selection contains only static columns.
* @return <code>true</code> if this selection contains only static columns, <code>false</code> otherwise;
*/
bool contains_only_static_columns() const;
/**
* Returns the index of the specified column, in the un-processed domain (before applying
* transformations to the input columns and aggregations)
*
* @param def the column definition
* @return the index of the specified column
*/
int32_t index_of(const column_definition& def) const;
bool has_column(const column_definition& def) const;
::shared_ptr<const metadata> get_result_metadata() const {
return _metadata;
}
::shared_ptr<metadata> get_result_metadata() {
return _metadata;
}
static ::shared_ptr<selection> wildcard(schema_ptr schema);
static std::vector<const column_definition*> wildcard_columns(schema_ptr schema);
static ::shared_ptr<selection> for_columns(schema_ptr schema, std::vector<const column_definition*> columns);
// Adds a column to the selection and result set. Returns an index within the result set row.
virtual uint32_t add_column_for_post_processing(const column_definition& c);
virtual std::vector<shared_ptr<functions::function>> used_functions() const { return {}; }
query::partition_slice::option_set get_query_options();
protected:
// Result of add_column: index in _columns and whether it was added now (or existed already).
struct add_column_result {
uint32_t index;
bool added;
};
// Adds a column to the _columns if not already present, returns add_column_result.
add_column_result add_column(const column_definition& c);
private:
static bool processes_selection(const std::vector<prepared_selector>& prepared_selectors);
static std::vector<lw_shared_ptr<column_specification>> collect_metadata(const schema& schema,
const std::vector<prepared_selector>& prepared_selectors);
public:
static ::shared_ptr<selection> from_selectors(data_dictionary::database db, schema_ptr schema, const sstring& ks, const std::vector<prepared_selector>& raw_selectors);
virtual std::unique_ptr<selectors> new_selectors() const = 0;
/**
* Returns a range of CQL3 columns this selection needs.
*/
auto const& get_columns() const {
return _columns;
}
uint32_t get_column_count() const {
return _columns.size();
}
virtual bool is_aggregate() const = 0;
virtual bool is_count() const {return false;}
virtual bool is_reducible() const {return false;}
virtual query::mapreduce_request::reductions_info get_reductions() const {return {{}, {}};}
/**
* Returns true if the selection is trivial, i.e. there are no function
* selectors (including casts or aggregates).
*/
bool is_trivial() const { return _is_trivial; }
friend class result_set_builder;
};
shared_ptr<selection> selection_from_partition_slice(schema_ptr schema, const query::partition_slice& slice);
class result_set_builder {
private:
std::unique_ptr<result_set> _result_set;
std::unique_ptr<selectors> _selectors;
const std::vector<size_t> _group_by_cell_indices; ///< Indices in \c current of cells holding GROUP BY values.
const uint64_t _limit; ///< Maximum number of rows to return.
const uint64_t _per_partition_limit; ///< Maximum number of rows to return per partition.
uint64_t _per_partition_remaining; ///< Remaining rows to return for the current partition.
uint64_t _per_partition_remaining_previous_partition; ///< Value of _per_partition_remaining before the current partition.
///< Necessary because the next page might continue the same partition,
///< but accept_partition_end() and accept_new_partition() will be called anyway.
std::vector<managed_bytes_opt> _last_group; ///< Previous row's group: all of GROUP BY column values.
bool _group_began; ///< Whether a group began being formed.
public:
std::vector<managed_bytes_opt> current;
std::vector<bytes> current_partition_key;
std::vector<bytes> current_clustering_key;
std::vector<api::timestamp_type> _timestamps;
std::vector<int32_t> _ttls;
const query_options* _options;
private:
const gc_clock::time_point _now;
public:
template<typename Func>
auto with_thread_if_needed(Func&& func) {
if (_selectors->requires_thread()) {
return async(std::move(func));
} else {
return futurize_invoke(std::move(func));
}
}
class nop_filter {
public:
inline bool operator()(const selection&, const std::vector<bytes>&, const std::vector<bytes>&, const query::result_row_view&, const query::result_row_view*) const {
return true;
}
void reset(const partition_key* = nullptr) {
}
uint64_t get_rows_dropped() const {
return 0;
}
};
class restrictions_filter {
const ::shared_ptr<const restrictions::statement_restrictions> _restrictions;
const query_options& _options;
const expr::expression& _partition_level_filter;
const expr::expression& _clustering_row_level_filter;
mutable bool _current_partition_does_not_match = false;
mutable uint64_t _rows_dropped = 0;
mutable uint64_t _remaining;
schema_ptr _schema;
mutable uint64_t _per_partition_limit;
mutable uint64_t _per_partition_remaining;
mutable uint64_t _rows_fetched_for_last_partition;
mutable std::optional<partition_key> _last_pkey;
mutable bool _is_first_partition_on_page = true;
public:
explicit restrictions_filter(::shared_ptr<const restrictions::statement_restrictions> restrictions,
const query_options& options,
uint64_t remaining,
schema_ptr schema,
uint64_t per_partition_limit,
std::optional<partition_key> last_pkey = {},
uint64_t rows_fetched_for_last_partition = 0);
bool operator()(const selection& selection, const std::vector<bytes>& pk, const std::vector<bytes>& ck, const query::result_row_view& static_row, const query::result_row_view* row) const;
void reset(const partition_key* key = nullptr);
uint64_t get_rows_dropped() const {
return _rows_dropped;
}
private:
bool do_filter(const selection& selection, const std::vector<bytes>& pk, const std::vector<bytes>& ck, const query::result_row_view& static_row, const query::result_row_view* row) const;
};
result_set_builder(const selection& s, gc_clock::time_point now,
const query_options* options = nullptr,
std::vector<size_t> group_by_cell_indices = {},
uint64_t limit = std::numeric_limits<uint64_t>::max(),
uint64_t per_partition_limit = std::numeric_limits<uint64_t>::max());
void add_empty();
void add(bytes_opt value);
void add(const column_definition& def, const query::result_atomic_cell_view& c);
void add_collection(const column_definition& def, bytes_view c);
void start_new_row();
void complete_row();
void accept_new_partition(const std::vector<bytes>& key);
void accept_partition_end();
std::unique_ptr<result_set> build();
api::timestamp_type timestamp_of(size_t idx);
int32_t ttl_of(size_t idx);
size_t result_set_size() const;
// Implements ResultVisitor concept from query.hh
template<typename Filter = nop_filter>
class visitor {
protected:
result_set_builder& _builder;
const schema& _schema;
const selection& _selection;
uint64_t _row_count;
std::vector<bytes>& _partition_key;
std::vector<bytes>& _clustering_key;
Filter _filter;
public:
visitor(cql3::selection::result_set_builder& builder, const schema& s,
const selection& selection, Filter filter = Filter())
: _builder(builder)
, _schema(s)
, _selection(selection)
, _row_count(0)
, _partition_key(_builder.current_partition_key)
, _clustering_key(_builder.current_clustering_key)
, _filter(filter)
{}
visitor(visitor&&) = default;
void add_value(const column_definition& def, query::result_row_view::iterator_type& i) {
if (def.type->is_multi_cell()) {
auto cell = i.next_collection_cell();
if (!cell) {
_builder.add_empty();
return;
}
_builder.add_collection(def, cell->linearize());
} else {
auto cell = i.next_atomic_cell();
if (!cell) {
_builder.add_empty();
return;
}
_builder.add(def, *cell);
}
}
void accept_new_partition(const partition_key& key, uint64_t row_count) {
auto partition_key = key.explode(_schema);
_builder.accept_new_partition(partition_key);
_partition_key = std::move(partition_key);
_row_count = row_count;
_filter.reset(&key);
}
void accept_new_partition(uint64_t row_count) {
_row_count = row_count;
_filter.reset();
}
void accept_new_row(const clustering_key& key, const query::result_row_view& static_row, const query::result_row_view& row) {
_clustering_key = key.explode(_schema);
accept_new_row(static_row, row);
}
void accept_new_row(const query::result_row_view& static_row, const query::result_row_view& row) {
auto static_row_iterator = static_row.iterator();
auto row_iterator = row.iterator();
if (!_filter(_selection, _partition_key, _clustering_key, static_row, &row)) {
return;
}
_builder.start_new_row();
for (auto&& def : _selection.get_columns()) {
switch (def->kind) {
case column_kind::partition_key:
_builder.add(_partition_key[def->component_index()]);
break;
case column_kind::clustering_key:
if (_clustering_key.size() > def->component_index()) {
_builder.add(_clustering_key[def->component_index()]);
} else {
_builder.add({});
}
break;
case column_kind::regular_column:
add_value(*def, row_iterator);
break;
case column_kind::static_column:
add_value(*def, static_row_iterator);
break;
default:
throwing_assert(0);
}
}
_builder.complete_row();
}
uint64_t accept_partition_end(const query::result_row_view& static_row) {
if (_row_count == 0) {
if (!_filter(_selection, _partition_key, _clustering_key, static_row, nullptr)) {
return _filter.get_rows_dropped();
}
_builder.start_new_row();
auto static_row_iterator = static_row.iterator();
for (auto&& def : _selection.get_columns()) {
if (def->is_partition_key()) {
_builder.add(_partition_key[def->component_index()]);
} else if (def->is_static()) {
add_value(*def, static_row_iterator);
} else {
_builder.add_empty();
}
}
_builder.complete_row();
}
_builder.accept_partition_end();
return _filter.get_rows_dropped();
}
};
private:
bytes_opt get_value(data_type t, query::result_atomic_cell_view c);
/// True iff the \c current row ends a previously started group, either according to
/// _group_by_cell_indices or aggregation.
bool last_group_ended() const;
/// Gets output row from _selectors and resets them.
void flush_selectors();
/// Updates _last_group from the \c current row.
void update_last_group();
};
}
}