cleanup(userspace/libsinsp): completely drop m_program_hash and m_program_hashscript fields from threadinfo.

They are unused and can be eventually directly implemented by consumers, if needed.

Signed-off-by: Federico Di Pierro <nierro92@gmail.com>

Author: FedeDP

userspace/libsinsp/parsers.cpp

@@ -2233,11 +2233,6 @@ void sinsp_parser::parse_execve_exit(sinsp_evt *evt) {
 	//
 	evt->get_tinfo()->m_flags |= PPM_CL_NAME_CHANGED;
 
-	//
-	// Recompute the program hash
-	//
-	evt->get_tinfo()->compute_program_hash();
-
 	//
 	// If there's a listener, add a callback to later invoke it.
 	//

userspace/libsinsp/threadinfo.cpp

@@ -100,8 +100,6 @@ libsinsp::state::static_struct::field_infos sinsp_threadinfo::static_fields() co
 	define_static_field(ret, this, m_pgid, "pgid");
 	define_static_field(ret, this, m_pidns_init_start_ts, "pidns_init_start_ts");
 	define_static_field(ret, this, m_root, "root");
-	// m_program_hash
-	// m_program_hash_scripts
 	define_static_field(ret, this, m_tty, "tty");
 	// m_category
 	// m_clone_ts
@@ -145,8 +143,6 @@ void sinsp_threadinfo::init() {
 	m_lastevent_fd = 0;
 	m_last_latency_entertime = 0;
 	m_latency = 0;
-	m_program_hash = 0;
-	m_program_hash_scripts = 0;
 	m_lastevent_data = NULL;
 	m_parent_loop_detected = false;
 	m_tty = 0;
@@ -203,57 +199,6 @@ void sinsp_threadinfo::fix_sockets_coming_from_proc() {
 	});
 }
 
-#define STR_AS_NUM_JAVA 0x6176616a
-#define STR_AS_NUM_RUBY 0x79627572
-#define STR_AS_NUM_PERL 0x6c726570
-#define STR_AS_NUM_NODE 0x65646f6e
-
-#define MAX_PROG_HASH_LEN 1024
-
-void sinsp_threadinfo::compute_program_hash() {
-	auto curr_hash = std::hash<std::string>()(m_exe);
-	hash_combine(curr_hash, m_container_id);
-	auto rem_len = MAX_PROG_HASH_LEN - (m_exe.size() + m_container_id.size());
-
-	//
-	// By default, the scripts hash is just exe+container
-	//
-	m_program_hash_scripts = curr_hash;
-
-	//
-	// The program hash includes the arguments as well
-	//
-	for(auto arg = m_args.begin(); arg != m_args.end() && rem_len > 0; ++arg) {
-		if(arg->size() >= rem_len) {
-			auto partial_str = arg->substr(0, rem_len);
-			hash_combine(curr_hash, partial_str);
-			break;
-		}
-
-		hash_combine(curr_hash, *arg);
-		rem_len -= arg->size();
-	}
-	m_program_hash = curr_hash;
-
-	//
-	// For some specific processes (essentially the scripting languages)
-	// we include the arguments in the scripts hash as well
-	//
-	if(m_comm.size() == 4) {
-		uint32_t ncomm;
-		memcpy(&ncomm, m_comm.c_str(), 4);
-
-		if(ncomm == STR_AS_NUM_JAVA || ncomm == STR_AS_NUM_RUBY || ncomm == STR_AS_NUM_PERL ||
-		   ncomm == STR_AS_NUM_NODE) {
-			m_program_hash_scripts = m_program_hash;
-		}
-	} else if(m_comm.size() >= 6) {
-		if(m_comm.substr(0, 6) == "python") {
-			m_program_hash_scripts = m_program_hash;
-		}
-	}
-}
-
 void sinsp_threadinfo::add_fd_from_scap(scap_fdinfo* fdi) {
 	auto newfdi = m_inspector->build_fdinfo();
 	bool do_add = true;
@@ -1477,7 +1422,6 @@ const std::shared_ptr<sinsp_threadinfo>& sinsp_thread_manager::add_thread(
 		        "adding entry with incompatible dynamic defs to of file descriptor sub-table");
 	}
 
-	tinfo_shared_ptr->compute_program_hash();
 	if(m_sinsp_stats_v2 != nullptr) {
 		m_sinsp_stats_v2->m_n_added_threads++;
 	}

userspace/libsinsp/threadinfo.h

@@ -434,10 +434,8 @@ class SINSP_PUBLIC sinsp_threadinfo : public libsinsp::state::table_entry {
 	int64_t m_pgid;   // Process group id, as seen from the host pid namespace
 	uint64_t m_pidns_init_start_ts;  ///< The pid_namespace init task (child_reaper) start_time ts.
 	std::string m_root;
-	size_t m_program_hash;          ///< Unique hash of the current program
-	size_t m_program_hash_scripts;  ///< Unique hash of the current program, including arguments for
-	                                ///< scripting programs (like python or ruby)
-	uint32_t m_tty;                 ///< Number of controlling terminal
+
+	uint32_t m_tty;  ///< Number of controlling terminal
 	std::shared_ptr<thread_group_info> m_tginfo;
 	std::list<std::weak_ptr<sinsp_threadinfo>> m_children;
 	uint64_t m_not_expired_children;
@@ -497,27 +495,6 @@ class SINSP_PUBLIC sinsp_threadinfo : public libsinsp::state::table_entry {
 	//
 	sinsp* m_inspector;
 
-	struct hasher {
-		size_t operator()(sinsp_threadinfo* tinfo) const {
-			auto main_thread = tinfo->get_main_thread();
-			if(main_thread == nullptr) {
-				return 0;
-			}
-			return main_thread->m_program_hash;
-		}
-	};
-
-	struct comparer {
-		size_t operator()(sinsp_threadinfo* lhs, sinsp_threadinfo* rhs) const {
-			auto lhs_main_thread = lhs->get_main_thread();
-			auto rhs_main_thread = rhs->get_main_thread();
-			if(lhs_main_thread == nullptr || rhs_main_thread == nullptr) {
-				return 0;
-			}
-			return lhs_main_thread->m_program_hash == rhs_main_thread->m_program_hash;
-		}
-	};
-
 	/* Note that `fd_table` should be shared with the main thread only if `PPM_CL_CLONE_FILES`
 	 * is specified. Today we always specify `PPM_CL_CLONE_FILES` for all threads.
 	 */
@@ -556,7 +533,6 @@ class SINSP_PUBLIC sinsp_threadinfo : public libsinsp::state::table_entry {
 			m_lastevent_cpuid = (uint16_t)-1;
 		}
 	}
-	void compute_program_hash();
 
 	inline const uint8_t* get_last_event_data() const { return m_lastevent_data; }