Initial support for lua transactions.

Extend multi-transactions to scripts. Differentiate between incremental and instant locking
for multi-transactions.
This commit is contained in:
Roman Gershman 2022-02-08 17:56:47 +02:00
parent b1829c3fe0
commit e1c852dfcc
10 changed files with 173 additions and 67 deletions

View File

@ -157,6 +157,18 @@ TEST_F(InterpreterTest, Basic) {
}
}
TEST_F(InterpreterTest, UnknownFunc) {
string_view code(R"(
function foo(n)
return myunknownfunc(1, n)
end)");
CHECK_EQ(0, luaL_loadbuffer(lua(), code.data(), code.size(), "code1"));
CHECK_EQ(0, lua_pcall(lua(), 0, 0, 0));
int type = lua_getglobal(lua(), "myunknownfunc");
ASSERT_EQ(LUA_TNIL, type);
}
TEST_F(InterpreterTest, Stack) {
RunInline(R"(
local x = {}

View File

@ -33,6 +33,7 @@ CommandRegistry::CommandRegistry() {
CommandId cd("COMMAND", CO::RANDOM | CO::LOADING, 0, 0, 0, 0);
cd.SetHandler([this](const auto& args, auto* cntx) { return Command(args, cntx); });
const char* nm = cd.name();
cmd_map_.emplace(nm, std::move(cd));
}
@ -70,6 +71,38 @@ CommandRegistry& CommandRegistry::operator<<(CommandId cmd) {
return *this;
}
KeyIndex DetermineKeys(const CommandId* cid, const CmdArgList& args) {
DCHECK_EQ(0u, cid->opt_mask() & CO::GLOBAL_TRANS);
KeyIndex key_index;
if (cid->first_key_pos() > 0) {
key_index.start = cid->first_key_pos();
int last = cid->last_key_pos();
key_index.end = last > 0 ? last + 1 : (int(args.size()) + 1 + last);
key_index.step = cid->key_arg_step();
return key_index;
}
string_view name{cid->name()};
if (name == "EVAL" || name == "EVALSHA") {
DCHECK_GE(args.size(), 3u);
uint32_t num_keys;
CHECK(absl::SimpleAtoi(ArgS(args, 2), &num_keys));
key_index.start = 3;
key_index.end = 3 + num_keys;
key_index.step = 1;
return key_index;
}
LOG(FATAL) << "TBD: Not supported";
return key_index;
}
namespace CO {
const char* OptName(CO::CommandOpt fl) {

View File

@ -37,8 +37,13 @@ const char* OptName(CommandOpt fl);
class CommandId {
public:
using Handler = std::function<void(CmdArgList, ConnectionContext*)>;
using ArgValidator = std::function<bool(CmdArgList, ConnectionContext*)>;
using Handler =
fu2::function_base<true /*owns*/, true /*copyable*/, fu2::capacity_default,
false /* non-throwing*/, false /* strong exceptions guarantees*/,
void(CmdArgList, ConnectionContext*) const>;
using ArgValidator = fu2::function_base<true, true, fu2::capacity_default, false, false,
bool(CmdArgList, ConnectionContext*) const>;
/**
* @brief Construct a new Command Id object
@ -152,4 +157,8 @@ class CommandRegistry {
void Command(CmdArgList args, ConnectionContext* cntx);
};
// Given the command and the arguments determines the keys range (index).
KeyIndex DetermineKeys(const CommandId* cid, const CmdArgList& args);
} // namespace dfly

View File

@ -43,6 +43,12 @@ struct KeyLockArgs {
unsigned key_step;
};
// Describes key indices.
struct KeyIndex {
unsigned start;
unsigned end; // does not include this index (open limit).
unsigned step; // 1 for commands like mget. 2 for commands like mset.
};
struct ConnectionStats {
uint32_t num_conns = 0;

View File

@ -4,6 +4,8 @@
#pragma once
#include <absl/container/flat_hash_set.h>
#include "server/common_types.h"
#include "server/reply_builder.h"
@ -50,6 +52,8 @@ struct ConnectionState {
// Lua-script related data.
struct Script {
bool is_write = true;
absl::flat_hash_set<std::string_view> keys;
};
std::optional<Script> script_info;
};

View File

@ -335,6 +335,9 @@ bool DbSlice::Acquire(IntentLock::Mode mode, const KeyLockArgs& lock_args) {
}
void DbSlice::Release(IntentLock::Mode mode, const KeyLockArgs& lock_args) {
DCHECK(!lock_args.args.empty());
DVLOG(2) << "Release " << IntentLock::ModeName(mode) << " for " << lock_args.args[0];
if (lock_args.args.size() == 1) {
Release(mode, lock_args.db_index, lock_args.args.front(), 1);
} else {
@ -351,7 +354,7 @@ void DbSlice::Release(IntentLock::Mode mode, const KeyLockArgs& lock_args) {
}
}
}
DVLOG(1) << "Release " << IntentLock::ModeName(mode) << " for " << lock_args.args[0];
}
}

View File

@ -221,8 +221,22 @@ TEST_F(DflyEngineTest, Eval) {
resp = Run({"incrby", "foo", "42"});
EXPECT_THAT(resp[0], IntArg(42));
// resp = Run({"eval", "return redis.call('get', 'foo')", "0"});
// EXPECT_THAT(resp[0], StrArg("42"));
resp = Run({"eval", "return redis.call('get', 'foo')", "0"});
EXPECT_THAT(resp[0], ErrArg("undeclared"));
resp = Run({"eval", "return redis.call('get', 'foo')", "1", "bar"});
EXPECT_THAT(resp[0], ErrArg("undeclared"));
ASSERT_FALSE(service_->IsLocked(0, "foo"));
resp = Run({"eval", "return redis.call('get', 'foo')", "1", "foo"});
EXPECT_THAT(resp[0], StrArg("42"));
resp = Run({"eval", "return redis.call('get', KEYS[1])", "1", "foo"});
EXPECT_THAT(resp[0], StrArg("42"));
ASSERT_FALSE(service_->IsLocked(0, "foo"));
ASSERT_FALSE(service_->IsShardSetLocked());
}
TEST_F(DflyEngineTest, EvalSha) {

View File

@ -430,7 +430,18 @@ void Service::DispatchCommand(CmdArgList args, ConnectionContext* cntx) {
// Create command transaction
intrusive_ptr<Transaction> dist_trans;
if (!under_script) {
if (under_script) {
DCHECK(cntx->transaction);
KeyIndex key_index = DetermineKeys(cid, args);
for (unsigned i = key_index.start; i < key_index.end; ++i) {
string_view key = ArgS(args, i);
if (!cntx->conn_state.script_info->keys.contains(key)) {
return (*cntx)->SendError("script tried accessing undeclared key");
}
}
cntx->transaction->SetExecCmd(cid);
cntx->transaction->InitByArgs(cntx->conn_state.db_index, args);
} else {
DCHECK(cntx->transaction == nullptr);
if (IsTransactional(cid)) {
@ -545,6 +556,7 @@ void Service::Multi(CmdArgList args, ConnectionContext* cntx) {
}
void Service::CallFromScript(CmdArgList args, ObjectExplorer* reply, ConnectionContext* cntx) {
DCHECK(cntx->transaction);
InterpreterReplier replier(reply);
ReplyBuilderInterface* orig = cntx->Inject(&replier);
@ -647,6 +659,13 @@ void Service::EvalInternal(const EvalArgs& eval_args, Interpreter* interpreter,
// and checking whether all invocations consist of RO commands.
// we can do it once during script insertion into script mgr.
cntx->conn_state.script_info.emplace();
for (size_t i = 0; i < eval_args.keys.size(); ++i) {
cntx->conn_state.script_info->keys.insert(ArgS(eval_args.keys, i));
}
DCHECK(cntx->transaction);
if (!eval_args.keys.empty())
cntx->transaction->Schedule();
auto lk = interpreter->Lock();
@ -659,6 +678,10 @@ void Service::EvalInternal(const EvalArgs& eval_args, Interpreter* interpreter,
cntx->conn_state.script_info.reset(); // reset script_info
// Conclude the transaction.
if (!eval_args.keys.empty())
cntx->transaction->UnlockMulti();
if (result == Interpreter::RUN_ERR) {
string resp = absl::StrCat("Error running script (call to ", eval_args.sha, "): ", error);
return (*cntx)->SendError(resp);

View File

@ -24,42 +24,6 @@ std::atomic_uint64_t op_seq{1};
[[maybe_unused]] constexpr size_t kTransSize = sizeof(Transaction);
struct KeyIndex {
unsigned start;
unsigned end; // not including
unsigned step;
};
KeyIndex DetermineKeys(const CommandId* cid, const CmdArgList& args) {
DCHECK_EQ(0u, cid->opt_mask() & CO::GLOBAL_TRANS);
KeyIndex key_index;
if (cid->first_key_pos() > 0) {
key_index.start = cid->first_key_pos();
int last = cid->last_key_pos();
key_index.end = last > 0 ? last + 1 : (int(args.size()) + 1 + last);
key_index.step = cid->key_arg_step();
return key_index;
}
string_view name{cid->name()};
if (name == "EVAL" || name == "EVALSHA") {
DCHECK_GE(args.size(), 3u);
uint32_t num_keys;
CHECK(absl::SimpleAtoi(ArgS(args, 2), &num_keys));
key_index.start = 3;
key_index.end = 3 + num_keys;
key_index.step = 1;
return key_index;
}
LOG(FATAL) << "Not supported";
return key_index;
}
} // namespace
@ -130,7 +94,7 @@ OpResult<Transaction::FindFirstResult> Transaction::FindFirstProcessor::Process(
}
IntentLock::Mode Transaction::Mode() const {
return (trans_options_ & CO::READONLY) ? IntentLock::SHARED : IntentLock::EXCLUSIVE;
return (cid_->opt_mask() & CO::READONLY) ? IntentLock::SHARED : IntentLock::EXCLUSIVE;
}
/**
@ -141,10 +105,15 @@ IntentLock::Mode Transaction::Mode() const {
* @param cs
*/
Transaction::Transaction(const CommandId* cid, EngineShardSet* ess) : cid_(cid), ess_(ess) {
if (strcmp(cid_->name(), "EXEC") == 0) {
string_view cmd_name(cid_->name());
if (cmd_name == "EXEC" || cmd_name == "EVAL" || cmd_name == "EVALSHA") {
multi_.reset(new Multi);
multi_->multi_opts = cid->opt_mask();
if (cmd_name == "EVAL" || cmd_name == "EVALSHA") {
multi_->incremental = false; // we lock all the keys at once.
}
}
trans_options_ = cid_->opt_mask();
}
Transaction::~Transaction() {
@ -185,22 +154,19 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
KeyIndex key_index = DetermineKeys(cid_, args);
if (key_index.start == args.size()) {
if (key_index.start == args.size()) { // eval with 0 keys.
CHECK(absl::StartsWith(cid_->name(), "EVAL"));
return;
}
DCHECK_LT(key_index.start, args.size());
DCHECK_GT(key_index.start, 0u);
bool single_key = key_index.start > 0 && !cid_->is_multi_key();
if (!multi_ && single_key) {
bool incremental_locking = multi_ && multi_->incremental;
bool single_key = !multi_ && (key_index.start + key_index.step) >= key_index.end;
if (single_key) {
shard_data_.resize(1); // Single key optimization
} else {
// Our shard_data is not sparse, so we must allocate for all threads :(
shard_data_.resize(ess_->size());
}
if (!multi_ && single_key) { // Single key optimization.
auto key = ArgS(args, key_index.start);
args_.push_back(key);
@ -210,10 +176,12 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
return;
}
// Our shard_data is not sparse, so we must allocate for all threads :(
shard_data_.resize(ess_->size());
CHECK(key_index.step == 1 || key_index.step == 2);
DCHECK(key_index.step == 1 || (args.size() % 2) == 1);
// Reuse thread-local temporary storage. Since this code is non-preemptive we can use it here.
// Reuse thread-local temporary storage. Since this code is atomic we can use it here.
auto& shard_index = tmp_space.shard_cache;
shard_index.resize(shard_data_.size());
for (auto& v : shard_index) {
@ -223,10 +191,12 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
// TODO: to determine correctly locking mode for transactions, scripts
// and regular commands.
IntentLock::Mode mode = IntentLock::EXCLUSIVE;
bool should_record_locks = false;
if (multi_) {
mode = Mode();
tmp_space.uniq_keys.clear();
DCHECK_LT(int(mode), 2);
should_record_locks = incremental_locking || !multi_->locks_recorded;
}
for (unsigned i = key_index.start; i < key_index.end; ++i) {
@ -235,7 +205,7 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
shard_index[sid].args.push_back(key);
shard_index[sid].original_index.push_back(i - 1);
if (multi_ && tmp_space.uniq_keys.insert(key).second) {
if (should_record_locks && tmp_space.uniq_keys.insert(key).second) {
multi_->locks[key].cnt[int(mode)]++;
};
@ -247,6 +217,10 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
}
}
if (multi_) {
multi_->locks_recorded = true;
}
args_.resize(key_index.end - key_index.start);
reverse_index_.resize(args_.size());
@ -263,7 +237,14 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
sd.arg_count = si.args.size();
sd.arg_start = next_arg - args_.begin();
sd.local_mask = 0;
// We reset the local_mask for incremental locking to allow locking of arguments
// for each operation within the same transaction. For instant locking we lock at
// the beginning all the keys so we must preserve the mask to avoid double locking.
if (incremental_locking) {
sd.local_mask = 0;
}
if (!sd.arg_count)
continue;
@ -297,7 +278,8 @@ void Transaction::InitByArgs(DbIndex index, CmdArgList args) {
// Validation.
for (const auto& sd : shard_data_) {
DCHECK_EQ(sd.local_mask, 0u);
// sd.local_mask may be non-zero for multi transactions with instant locking.
// Specifically EVALs may maintain state between calls.
DCHECK_EQ(0, sd.local_mask & ARMED);
if (!multi_) {
DCHECK_EQ(TxQueue::kEnd, sd.pq_pos);
@ -309,13 +291,17 @@ void Transaction::SetExecCmd(const CommandId* cid) {
DCHECK(multi_);
DCHECK(!cb_);
// The order is important, we are Schedule() for multi transaction before overriding cid_.
// TODO: The flow is ugly. I should introduce a proper interface for Multi transactions
// like SetupMulti/TurndownMulti. We already have UnlockMulti that should be part of
// TurndownMulti.
if (txid_ == 0) {
Schedule();
}
unique_shard_cnt_ = 0;
cid_ = cid;
trans_options_ = cid->opt_mask();
cb_ = nullptr;
}
@ -344,6 +330,7 @@ bool Transaction::RunInShard(EngineShard* shard) {
DCHECK_EQ(sd.local_mask & (SUSPENDED_Q | EXPIRED_Q), 0);
bool awaked_prerun = (sd.local_mask & AWAKED_Q) != 0;
bool incremental_lock = multi_ && multi_->incremental;
// For multi we unlock transaction (i.e. its keys) in UnlockMulti() call.
// Therefore we differentiate between concluding, which says that this specific
@ -354,8 +341,8 @@ bool Transaction::RunInShard(EngineShard* shard) {
IntentLock::Mode mode = Mode();
// We make sure that we lock exactly once for each (multi-hop) transaction inside
// multi-transactions.
if (multi_ && ((sd.local_mask & KEYLOCK_ACQUIRED) == 0)) {
// transactions that lock incrementally.
if (incremental_lock && ((sd.local_mask & KEYLOCK_ACQUIRED) == 0)) {
DCHECK(!awaked_prerun); // we should not have blocking transaction inside multi block.
sd.local_mask |= KEYLOCK_ACQUIRED;
@ -458,6 +445,11 @@ void Transaction::ScheduleInternal() {
uint32_t num_shards;
std::function<bool(uint32_t)> is_active;
// TODO: For multi-transactions we should be able to deduce mode() at run-time based
// on the context. For regular multi-transactions we can actually inspect all commands.
// For eval-like transactions - we can decided based on the command flavor (EVAL/EVALRO) or
// auto-tune based on the static analysis (by identifying commands with hardcoded command names).
IntentLock::Mode mode = Mode();
if (span_all) {
@ -534,6 +526,10 @@ OpStatus Transaction::ScheduleSingleHop(RunnableType cb) {
// single hop -> concluding.
coordinator_state_ |= (COORD_EXEC | COORD_EXEC_CONCLUDING);
if (!multi_) { // for non-multi transactions we schedule exactly once.
DCHECK_EQ(0, coordinator_state_ & COORD_SCHED);
}
bool schedule_fast = (unique_shard_cnt_ == 1) && !IsGlobal() && !multi_;
if (schedule_fast) { // Single shard (local) optimization.
// We never resize shard_data because that would affect MULTI transaction correctness.
@ -598,7 +594,10 @@ void Transaction::UnlockMulti() {
auto cb = [&] {
EngineShard* shard = EngineShard::tlocal();
shard->shard_lock()->Release(IntentLock::EXCLUSIVE);
if (multi_->multi_opts & CO::GLOBAL_TRANS) {
shard->shard_lock()->Release(IntentLock::EXCLUSIVE);
}
ShardId sid = shard->shard_id();
for (const auto& k_v : sharded_keys[sid]) {
auto release = [&](IntentLock::Mode mode) {
@ -606,6 +605,7 @@ void Transaction::UnlockMulti() {
shard->db_slice().Release(mode, this->db_index_, k_v.first, k_v.second.cnt[mode]);
}
};
release(IntentLock::SHARED);
release(IntentLock::EXCLUSIVE);
}
@ -1102,7 +1102,7 @@ inline uint32_t Transaction::DecreaseRunCnt() {
}
bool Transaction::IsGlobal() const {
return (trans_options_ & CO::GLOBAL_TRANS) != 0;
return (cid_->opt_mask() & CO::GLOBAL_TRANS) != 0;
}
// Runs only in the shard thread.

View File

@ -127,8 +127,7 @@ class Transaction {
return txid_;
}
// TODO: for multi trans_options_ changes with every operation.
// Does it mean we lock every key differently during the same transaction?
// based on cid_->opt_mask.
IntentLock::Mode Mode() const;
const char* Name() const;
@ -266,6 +265,10 @@ class Transaction {
struct Multi {
absl::flat_hash_map<std::string_view, LockCnt> locks;
uint32_t multi_opts = 0; // options of the parent transaction.
bool incremental = true;
bool locks_recorded = false;
};
util::fibers_ext::EventCount blocking_ec_; // used to wake blocking transactions.
@ -298,8 +301,6 @@ class Transaction {
// unique_shard_cnt_ and unique_shard_id_ is accessed only by coordinator thread.
uint32_t unique_shard_cnt_{0}; // number of unique shards span by args_
uint32_t trans_options_ = 0;
ShardId unique_shard_id_{kInvalidSid};
DbIndex db_index_ = 0;
@ -309,6 +310,7 @@ class Transaction {
enum CoordinatorState : uint8_t {
COORD_SCHED = 1,
COORD_EXEC = 2,
// We are running the last execution step in multi-hop operation.
COORD_EXEC_CONCLUDING = 4,
COORD_BLOCKED = 8,