commit 49b60a3: [Project] (Re)implement hyperscan caching
Vsevolod Stakhov
vsevolod at rspamd.com
Sun Oct 16 14:28:05 UTC 2022
Author: Vsevolod Stakhov
Date: 2022-10-16 15:22:27 +0100
URL: https://github.com/rspamd/rspamd/commit/49b60a37470e59aeadc7c25a966a656d054a2bc2
[Project] (Re)implement hyperscan caching
---
src/libutil/cxx/hyperscan_tools.cxx | 200 +++++++++++++++++++++++++++++++++++-
1 file changed, 197 insertions(+), 3 deletions(-)
diff --git a/src/libutil/cxx/hyperscan_tools.cxx b/src/libutil/cxx/hyperscan_tools.cxx
index 82664d810..f5086d1e4 100644
--- a/src/libutil/cxx/hyperscan_tools.cxx
+++ b/src/libutil/cxx/hyperscan_tools.cxx
@@ -20,10 +20,27 @@
#include "contrib/ankerl/unordered_dense.h"
#include "contrib/ankerl/svector.h"
#include "fmt/core.h"
+#include "locked_file.hxx"
+#include "hs.h"
+#include "logger.h"
+#include "locked_file.hxx"
#include <glob.h> /* for glob */
-#include <sys/stat.h> /* for stat */
#include <unistd.h> /* for unlink */
+#include <optional>
+#include "unix-std.h"
+
+#define msg_info_hyperscan(...) rspamd_default_log_function (G_LOG_LEVEL_INFO, \
+ "hyperscan", "", \
+ RSPAMD_LOG_FUNC, \
+ __VA_ARGS__)
+
+#define msg_debug_hyperscan(...) rspamd_conditional_debug_fast (NULL, NULL, \
+ rspamd_hyperscan_log_id, "hyperscan", "", \
+ RSPAMD_LOG_FUNC, \
+ __VA_ARGS__)
+
+INIT_LOG_MODULE_PUBLIC(hyperscan)
namespace rspamd::util {
@@ -45,7 +62,7 @@ private:
glob_t globbuf;
auto glob_pattern = fmt::format("{}{}*.{}",
- dir, G_DIR_SEPARATOR_S, ext);
+ dir, G_DIR_SEPARATOR_S, ext);
memset(&globbuf, 0, sizeof(globbuf));
if (glob(glob_pattern.c_str(), 0, nullptr, &globbuf) == 0) {
@@ -54,13 +71,20 @@ private:
struct stat st;
if (stat(path, &st) == -1) {
+ msg_debug_hyperscan("cannot stat file %s: %s",
+ path, strerror(errno));
continue;
}
if (S_ISREG(st.st_mode)) {
if (!known_cached_files.contains(path)) {
+ msg_info_hyperscan("remove stale hyperscan file %s", path);
unlink(path);
}
+ else {
+ msg_debug_hyperscan("found known hyperscan file %s, size: %Hz",
+ path, st.st_size);
+ }
}
}
}
@@ -88,10 +112,180 @@ public:
return *singleton;
}
+
+ void add_cached_file(const raii_file &file) {
+ auto dir = file.get_dir();
+ auto ext = file.get_extension();
+
+ if (std::find_if(cache_dirs.begin(), cache_dirs.end(),
+ [&](const auto& item){ return item == dir; }) == std::end(cache_dirs)) {
+ cache_dirs.emplace_back(std::string{dir});
+ }
+ if (std::find_if(cache_extensions.begin(), cache_extensions.end(),
+ [&](const auto& item){ return item == ext; }) == std::end(cache_extensions)) {
+ cache_extensions.emplace_back(std::string{ext});
+ }
+
+ known_cached_files.insert(file.get_name());
+ msg_debug_hyperscan("added new known hyperscan file: %*s", (int)file.get_name().size(),
+ file.get_name().data());
+ }
};
+/**
+ * This is a higher level representation of the cached hyperscan file
+ */
+struct hs_shared_database {
+ hs_database_t *db; /**< internal database (might be in a shared memory) */
+ std::optional<raii_mmaped_file> maybe_map;
+
+ ~hs_shared_database() {
+ if (!maybe_map) {
+ hs_free_database(db);
+ }
+ // Otherwise, handled by maybe_map dtor
+ }
+
+ explicit hs_shared_database(raii_mmaped_file &&map, hs_database_t *db) : db(db), maybe_map(std::move(map)) {}
+ explicit hs_shared_database(hs_database_t *db) : db(db), maybe_map(std::nullopt) {}
+};
+
+static auto
+hs_shared_from_unserialized(raii_mmaped_file &&map) -> tl::expected<hs_shared_database, std::string>
+{
+ auto ptr = map.get_map();
+ return tl::expected<hs_shared_database, std::string>{tl::in_place, std::move(map), (hs_database_t *)ptr};
+}
+
+static auto
+hs_shared_from_serialized(raii_mmaped_file &&map) -> tl::expected<hs_shared_database, std::string>
+{
+ hs_database_t *target = nullptr;
+
+ if (auto ret = hs_deserialize_database((const char *)map.get_map(), map.get_size(), &target); ret != HS_SUCCESS) {
+ return tl::make_unexpected("cannot deserialize database");
+ }
+
+ return tl::expected<hs_shared_database, std::string>{tl::in_place, target};
+}
+
+auto load_cached_hs_file(const char *fname, bool need_cleanup) -> tl::expected<hs_shared_database, std::string>
+{
+ auto &hs_cache = hs_known_files_cache::get(need_cleanup);
+
+ return raii_mmaped_file::mmap_shared(fname, O_RDONLY, PROT_READ)
+ .and_then([&]<class T>(T &&cached_serialized) -> tl::expected<hs_shared_database, std::string> {
+#if defined(HS_MAJOR) && defined(HS_MINOR) && HS_MAJOR >= 5 && HS_MINOR >= 4
+ auto unserialized_fname = fmt::format("{}.unser", fname);
+ auto unserialized_file = raii_locked_file::create(unserialized_fname.c_str(), O_CREAT | O_RDWR | O_EXCL,
+ 00644)
+ .and_then([&](auto &&new_file_locked) -> tl::expected<raii_file, std::string> {
+ auto tmpfile_pattern = fmt::format("{}{}hsmp-XXXXXXXXXXXXXXXXXX",
+ cached_serialized.get_file().get_dir(), G_DIR_SEPARATOR);
+ auto tmpfile = raii_locked_file::mkstemp(tmpfile_pattern.data(), O_CREAT | O_RDWR | O_EXCL,
+ 00644);
+
+ if (!tmpfile) {
+ return tl::make_unexpected(tmpfile.error());
+ }
+ else {
+ auto &tmpfile_checked = tmpfile.value();
+ std::size_t unserialized_size;
+
+ hs_serialized_database_size((const char *)cached_serialized.get_map(),
+ cached_serialized.get_size(), &unserialized_size);
+
+ msg_debug("multipattern: create new database in %s; %Hz size",
+ tmpfile_pattern.data(), unserialized_size);
+ void *buf;
+ posix_memalign(&buf, 16, unserialized_size);
+ if (buf == NULL) {
+ return tl::make_unexpected("Cannot allocate memory");
+ }
+
+ // Store owned string
+ auto tmpfile_name = std::string{tmpfile_checked.get_name()};
+
+ if (auto ret = hs_deserialize_database_at((const char *)cached_serialized.get_map(),
+ cached_serialized.get_size(), (hs_database_t *) buf); ret != HS_SUCCESS) {
+ return tl::make_unexpected(
+ fmt::format("cannot deserialize hyperscan database: {}", ret));
+ }
+ else {
+ if (write(tmpfile_checked.get_fd(), buf, unserialized_size) == -1) {
+ free(buf);
+ return tl::make_unexpected(fmt::format("cannot write to {}: {}",
+ tmpfile_name, ::strerror(errno)));
+ }
+ else {
+ free(buf);
+
+ /*
+ * Unlink target file before renaming to avoid
+ * race condition.
+ * So what we have is that `new_file_locked`
+ * will have flock on that file, so it will be
+ * replaced after unlink safely, and also unlocked.
+ */
+ (void) unlink(unserialized_fname.c_str());
+ if (rename(tmpfile_name.c_str(),
+ unserialized_fname.c_str()) == -1) {
+ if (errno != EEXIST) {
+ msg_err("cannot rename %s -> %s: %s",
+ tmpfile_name.c_str(),
+ unserialized_fname.c_str(),
+ strerror(errno));
+ }
+ }
+ else {
+ /* Unlock file but mark it as immortal first to avoid deletion */
+ tmpfile_checked.make_immortal();
+ (void) tmpfile_checked.unlock();
+ }
+ }
+ }
+ /* Reopen in RO mode */
+ return raii_file::open(unserialized_fname.c_str(), O_RDONLY);
+ };
+ })
+ .or_else([&](auto unused) -> tl::expected<raii_file, std::string> {
+ // Cannot create file, so try to open it in RO mode
+ return raii_file::open(unserialized_fname.c_str(), O_RDONLY);
+ });
+
+ hs_cache.add_cached_file(cached_serialized.get_file());
+
+ if (unserialized_file.has_value()) {
+
+ auto &unserialized_checked = unserialized_file.value();
+ hs_cache.add_cached_file(unserialized_checked);
+
+ if (unserialized_checked.get_size() == 0) {
+ /*
+ * This is a case when we have a file that is currently
+ * being created by another process.
+ * We cannot use it!
+ */
+ return hs_shared_from_serialized(std::forward<T>(cached_serialized));
+ }
+ else {
+ return raii_mmaped_file::mmap_shared(std::move(unserialized_checked), PROT_READ)
+ .and_then([&]<class U>(U &&mmapped_unserialized) -> auto {
+ return hs_shared_from_unserialized(std::forward<U>(mmapped_unserialized));
+ });
+ }
+ }
+ else {
+ return hs_shared_from_serialized(std::forward<T>(cached_serialized));
+ }
+#else // defined(HS_MAJOR) && defined(HS_MINOR) && HS_MAJOR >= 5 && HS_MINOR >= 4
+ hs_cache.add_cached_file(cached_serialized.get_file());
+ return hs_shared_from_serialized(std::forward<T>(cached_serialized));
+#endif // defined(HS_MAJOR) && defined(HS_MINOR) && HS_MAJOR >= 5 && HS_MINOR >= 4
+ });
+}
} // namespace rspamd::util
-#endif
\ No newline at end of file
+#endif // WITH_HYPERSCAN
\ No newline at end of file
More information about the Commits
mailing list