Remove long-obsolete members from the state file.

Tor has a feature to preserve unrecognized state file entries in
order to maintain forward compatibility.  But this feature, along
with some unused code that we never actually removed, led to us
keeping items that were of no use to the user, other than at worst
to preserve ancient information about them.

This commit adds a feature to remove obsolete entries when we load
the file.

Closes ticket 40137.

3 files changed, 62 insertions, 24 deletions

diff --git a/src/app/config/or_state_st.h b/src/app/config/or_state_st.h
index 31b7f8a983..6769ef7b87 100644
--- a/src/app/config/or_state_st.h
+++ b/src/app/config/or_state_st.h
@@ -38,17 +38,11 @@ struct or_state_t {
   uint64_t AccountingBytesAtSoftLimit;
   uint64_t AccountingExpectedUsage;
 
-  /** A list of Entry Guard-related configuration lines. (pre-prop271) */
-  struct config_line_t *EntryGuards;
-
-  /** A list of guard-related configuration lines. (post-prop271) */
+  /** A list of guard-related configuration lines. */
   struct config_line_t *Guard;
 
   struct config_line_t *TransportProxies;
 
-  /** Cached revision counters for active hidden services on this host */
-  struct config_line_t *HidServRevCounter;
-
   /** These fields hold information on the history of bandwidth usage for
    * servers.  The "Ends" fields hold the time when we last updated the
    * bandwidth usage. The "Interval" fields hold the granularity, in seconds,
diff --git a/src/app/config/statefile.c b/src/app/config/statefile.c
index b25167d2ec..22b15fcf24 100644
--- a/src/app/config/statefile.c
+++ b/src/app/config/statefile.c
@@ -58,16 +58,38 @@
 
 /** A list of state-file "abbreviations," for compatibility. */
 static config_abbrev_t state_abbrevs_[] = {
-  { "AccountingBytesReadInterval", "AccountingBytesReadInInterval", 0, 0 },
-  { "HelperNode", "EntryGuard", 0, 0 },
-  { "HelperNodeDownSince", "EntryGuardDownSince", 0, 0 },
-  { "HelperNodeUnlistedSince", "EntryGuardUnlistedSince", 0, 0 },
-  { "EntryNode", "EntryGuard", 0, 0 },
-  { "EntryNodeDownSince", "EntryGuardDownSince", 0, 0 },
-  { "EntryNodeUnlistedSince", "EntryGuardUnlistedSince", 0, 0 },
   { NULL, NULL, 0, 0},
 };
 
+/** A list of obsolete keys that we do not and should not preserve.
+ *
+ * We could just let these live in ExtraLines indefinitely, but they're
+ * never going to be used again, and every version that used them
+ * has been obsolete for a long time.
+ * */
+static const char *obsolete_state_keys[] = {
+  /* These were renamed in 0.1.1.11-alpha */
+  "AccountingBytesReadInterval",
+  "HelperNode",
+  "HelperNodeDownSince",
+  "HelperNodeUnlistedSince",
+  "EntryNode",
+  "HelperNodeDownSince",
+  "EntryNodeUnlistedSince",
+  /* These were replaced by "Guard" in 0.3.0.1-alpha. */
+  "EntryGuard",
+  "EntryGuardDownSince",
+  "EntryGuardUnlistedSince",
+  "EntryGuardAddedBy",
+  "EntryGuardPathBias",
+  "EntryGuardPathUseBias",
+  /* This was replaced by OPE-based revision numbers in 0.3.5.1-alpha,
+   * and was never actually used in a released version. */
+  "HidServRevCounter",
+
+  NULL,
+};
+
 /** dummy instance of or_state_t, used for type-checking its
  * members with CONF_CHECK_VAR_TYPE. */
 DUMMY_TYPECHECK_INSTANCE(or_state_t);
@@ -91,19 +113,9 @@ static const config_var_t state_vars_[] = {
   V(AccountingSoftLimitHitAt,         ISOTIME,  NULL),
   V(AccountingBytesAtSoftLimit,       MEMUNIT,  NULL),
 
-  VAR("EntryGuard",              LINELIST_S,  EntryGuards,             NULL),
-  VAR("EntryGuardDownSince",     LINELIST_S,  EntryGuards,             NULL),
-  VAR("EntryGuardUnlistedSince", LINELIST_S,  EntryGuards,             NULL),
-  VAR("EntryGuardAddedBy",       LINELIST_S,  EntryGuards,             NULL),
-  VAR("EntryGuardPathBias",      LINELIST_S,  EntryGuards,             NULL),
-  VAR("EntryGuardPathUseBias",   LINELIST_S,  EntryGuards,             NULL),
-  V(EntryGuards,                 LINELIST_V,  NULL),
-
   VAR("TransportProxy",               LINELIST_S, TransportProxies, NULL),
   V(TransportProxies,                 LINELIST_V, NULL),
 
-  V(HidServRevCounter,            LINELIST, NULL),
-
   V(BWHistoryReadEnds,                ISOTIME,  NULL),
   V(BWHistoryReadInterval,            POSINT,     "900"),
   V(BWHistoryReadValues,              CSV,      ""),
@@ -475,6 +487,7 @@ or_state_load(void)
   } else {
     log_info(LD_GENERAL, "Initialized state");
   }
+  or_state_remove_obsolete_lines(&new_state->ExtraLines);
   if (or_state_set(new_state) == -1) {
     or_state_save_broken(fname);
   }
@@ -494,6 +507,36 @@ or_state_load(void)
   return r;
 }
 
+/** Remove from `extra_lines` every element whose key appears in
+ * `obsolete_state_keys`. */
+STATIC void
+or_state_remove_obsolete_lines(config_line_t **extra_lines)
+{
+  /* make a strmap for the obsolete state names, so we can have O(1)
+     lookup. */
+  strmap_t *bad_keys = strmap_new();
+  for (unsigned i = 0; obsolete_state_keys[i] != NULL; ++i) {
+    strmap_set_lc(bad_keys, obsolete_state_keys[i], (void*)"rmv");
+  }
+
+  config_line_t **line = extra_lines;
+  while (*line) {
+    if (strmap_get_lc(bad_keys, (*line)->key) != NULL) {
+      /* This key is obsolete; remove it. */
+      config_line_t *victim = *line;
+      *line = (*line)->next;
+
+      victim->next = NULL; // prevent double-free.
+      config_free_lines(victim);
+    } else {
+      /* This is just an unrecognized key; keep it. */
+      line = &(*line)->next;
+    }
+  }
+
+  strmap_free(bad_keys, NULL);
+}
+
 /** Did the last time we tried to write the state file fail? If so, we
  * should consider disabling such features as preemptive circuit generation
  * to compute circuit-build-time. */
diff --git a/src/app/config/statefile.h b/src/app/config/statefile.h
index 98d9d2dda1..89b10560f3 100644
--- a/src/app/config/statefile.h
+++ b/src/app/config/statefile.h
@@ -33,6 +33,7 @@ STATIC void or_state_free_(or_state_t *state);
 STATIC or_state_t *or_state_new(void);
 struct config_mgr_t;
 STATIC const struct config_mgr_t *get_state_mgr(void);
+STATIC void or_state_remove_obsolete_lines(struct config_line_t **extra_lines);
 #endif /* defined(STATEFILE_PRIVATE) */
 
 #endif /* !defined(TOR_STATEFILE_H) */