diff options
author | Jakub Hrozek <jhrozek@redhat.com> | 2011-04-29 15:59:17 +0200 |
---|---|---|
committer | Stephen Gallagher <sgallagh@redhat.com> | 2011-05-20 07:21:45 -0400 |
commit | 7bdaf2a712d73763e7c3d25f6bb544b18f7028eb (patch) | |
tree | dca11da1c2c4b258d4bc33aed583d9ba48147f27 /src | |
parent | 6b95a91c1a49c2eff480820cfd8be51d70a29ffe (diff) | |
download | sssd2-7bdaf2a712d73763e7c3d25f6bb544b18f7028eb.tar.gz sssd2-7bdaf2a712d73763e7c3d25f6bb544b18f7028eb.tar.xz sssd2-7bdaf2a712d73763e7c3d25f6bb544b18f7028eb.zip |
Use dereference when processing RFC2307bis nested groups
Instead of issuing N LDAP requests when processing a group with N users,
utilize the dereference functionality to pull down all the members in a
single LDAP request.
https://fedorahosted.org/sssd/ticket/799
Diffstat (limited to 'src')
-rw-r--r-- | src/config/SSSDConfig.py | 1 | ||||
-rw-r--r-- | src/config/etc/sssd.api.d/sssd-ipa.conf | 2 | ||||
-rw-r--r-- | src/config/etc/sssd.api.d/sssd-ldap.conf | 1 | ||||
-rw-r--r-- | src/man/sssd-ldap.5.xml | 23 | ||||
-rw-r--r-- | src/providers/ipa/ipa_common.c | 3 | ||||
-rw-r--r-- | src/providers/ipa/ipa_common.h | 2 | ||||
-rw-r--r-- | src/providers/ldap/ldap_common.c | 3 | ||||
-rw-r--r-- | src/providers/ldap/sdap.h | 1 | ||||
-rw-r--r-- | src/providers/ldap/sdap_async_accounts.c | 468 |
9 files changed, 487 insertions, 17 deletions
diff --git a/src/config/SSSDConfig.py b/src/config/SSSDConfig.py index 5a26e31a..5114a178 100644 --- a/src/config/SSSDConfig.py +++ b/src/config/SSSDConfig.py @@ -145,6 +145,7 @@ option_strings = { 'ldap_deref' : _('How to dereference aliases'), 'ldap_dns_service_name' : _('Service name for DNS service lookups'), 'ldap_page_size' : _('The number of records to retrieve in a single LDAP query'), + 'ldap_deref_threshold' : _('The number of members that must be missing to trigger a full deref'), 'ldap_entry_usn' : _('entryUSN attribute'), 'ldap_rootdse_last_usn' : _('lastUSN attribute'), diff --git a/src/config/etc/sssd.api.d/sssd-ipa.conf b/src/config/etc/sssd.api.d/sssd-ipa.conf index 7719069c..75222250 100644 --- a/src/config/etc/sssd.api.d/sssd-ipa.conf +++ b/src/config/etc/sssd.api.d/sssd-ipa.conf @@ -35,6 +35,8 @@ ldap_referrals = bool, None, false ldap_krb5_ticket_lifetime = int, None, false ldap_dns_service_name = str, None, false ldap_deref = str, None, false +ldap_page_size = int, None, false +ldap_deref_threshold = int, None, false [provider/ipa/id] ldap_search_timeout = int, None, false diff --git a/src/config/etc/sssd.api.d/sssd-ldap.conf b/src/config/etc/sssd.api.d/sssd-ldap.conf index e568c74d..ce9ec513 100644 --- a/src/config/etc/sssd.api.d/sssd-ldap.conf +++ b/src/config/etc/sssd.api.d/sssd-ldap.conf @@ -28,6 +28,7 @@ ldap_krb5_ticket_lifetime = int, None, false ldap_dns_service_name = str, None, false ldap_deref = str, None, false ldap_page_size = int, None, false +ldap_deref_threshold = int, None, false [provider/ldap/id] ldap_search_timeout = int, None, false diff --git a/src/man/sssd-ldap.5.xml b/src/man/sssd-ldap.5.xml index 49c9e491..42ea848a 100644 --- a/src/man/sssd-ldap.5.xml +++ b/src/man/sssd-ldap.5.xml @@ -869,6 +869,29 @@ </varlistentry> <varlistentry> + <term>ldap_deref_threshold (integer)</term> + <listitem> + <para> + Specify the number of group members that must be + missing from the internal cache in order to trigger + a dereference lookup. If less members are missing, + they are looked up individually. + </para> + <para> + A dereference lookup is a means of fetching all + group members in a single LDAP call. + Different LDAP servers may implement different + dereference methods. The currently supported + servers are 389/RHDS, OpenLDAP and Active + Directory. + </para> + <para> + Default: 10 + </para> + </listitem> + </varlistentry> + + <varlistentry> <term>ldap_tls_reqcert (string)</term> <listitem> <para> diff --git a/src/providers/ipa/ipa_common.c b/src/providers/ipa/ipa_common.c index b0fe523c..da3fcd6f 100644 --- a/src/providers/ipa/ipa_common.c +++ b/src/providers/ipa/ipa_common.c @@ -94,7 +94,8 @@ struct dp_option ipa_def_ldap_opts[] = { * manpages or SSSDConfig API */ { "ldap_auth_disable_tls_never_use_in_production", DP_OPT_BOOL, BOOL_FALSE, BOOL_FALSE }, - { "ldap_page_size", DP_OPT_NUMBER, { .number = 1000 }, NULL_NUMBER } + { "ldap_page_size", DP_OPT_NUMBER, { .number = 1000 }, NULL_NUMBER }, + { "ldap_deref_threshold", DP_OPT_NUMBER, { .number = 10 }, NULL_NUMBER } }; struct sdap_attr_map ipa_attr_map[] = { diff --git a/src/providers/ipa/ipa_common.h b/src/providers/ipa/ipa_common.h index 02c4276c..8f0f35a8 100644 --- a/src/providers/ipa/ipa_common.h +++ b/src/providers/ipa/ipa_common.h @@ -35,7 +35,7 @@ struct ipa_service { /* the following defines are used to keep track of the options in the ldap * module, so that if they change and ipa is not updated correspondingly * this will trigger a runtime abort error */ -#define IPA_OPTS_BASIC_TEST 50 +#define IPA_OPTS_BASIC_TEST 51 /* the following define is used to keep track of the options in the krb5 * module, so that if they change and ipa is not updated correspondingly diff --git a/src/providers/ldap/ldap_common.c b/src/providers/ldap/ldap_common.c index 12028b01..1b52190f 100644 --- a/src/providers/ldap/ldap_common.c +++ b/src/providers/ldap/ldap_common.c @@ -87,7 +87,8 @@ struct dp_option default_basic_opts[] = { * manpages or SSSDConfig API */ { "ldap_auth_disable_tls_never_use_in_production", DP_OPT_BOOL, BOOL_FALSE, BOOL_FALSE }, - { "ldap_page_size", DP_OPT_NUMBER, { .number = 1000 }, NULL_NUMBER } + { "ldap_page_size", DP_OPT_NUMBER, { .number = 1000 }, NULL_NUMBER }, + { "ldap_deref_threshold", DP_OPT_NUMBER, { .number = 10 }, NULL_NUMBER } }; struct sdap_attr_map generic_attr_map[] = { diff --git a/src/providers/ldap/sdap.h b/src/providers/ldap/sdap.h index e98b27bc..4b662c15 100644 --- a/src/providers/ldap/sdap.h +++ b/src/providers/ldap/sdap.h @@ -194,6 +194,7 @@ enum sdap_basic_opt { SDAP_ENUM_SEARCH_TIMEOUT, SDAP_DISABLE_AUTH_TLS, SDAP_PAGE_SIZE, + SDAP_DEREF_THRESHOLD, SDAP_OPTS_BASIC /* opts counter */ }; diff --git a/src/providers/ldap/sdap_async_accounts.c b/src/providers/ldap/sdap_async_accounts.c index 89feff05..720ce96c 100644 --- a/src/providers/ldap/sdap_async_accounts.c +++ b/src/providers/ldap/sdap_async_accounts.c @@ -3278,6 +3278,28 @@ int sdap_get_initgr_recv(struct tevent_req *req) return EOK; } +struct sdap_deref_ctx { + const char *orig_dn; + + size_t expired_users_num; + uint32_t expired_users_index; + char **expired_users; + + size_t expired_groups_num; + uint32_t expired_groups_index; + char **expired_groups; + + size_t missing_dns_num; + uint32_t missing_dns_index; + char **missing_dns; + + struct sdap_deref_attrs **deref_result; + size_t num_results; + uint32_t result_index; + + int deref_threshold; +}; + struct sdap_nested_group_ctx { struct tevent_context *ev; struct sysdb_ctx *sysdb; @@ -3294,9 +3316,13 @@ struct sdap_nested_group_ctx { struct ldb_message_element *members; uint32_t member_index; char *member_dn; + + struct sdap_deref_ctx *derefctx; }; +static errno_t sdap_nested_group_process_deref_step(struct tevent_req *req); static errno_t sdap_nested_group_process_step(struct tevent_req *req); + static struct tevent_req *sdap_nested_group_process_send( TALLOC_CTX *mem_ctx, struct tevent_context *ev, struct sss_domain_info *domain, @@ -3390,8 +3416,20 @@ static struct tevent_req *sdap_nested_group_process_send( state->member_index = 0; - ret = sdap_nested_group_process_step(req); - if (ret != EAGAIN) goto immediate; + if (sdap_has_deref_support(state->sh)) { + state->derefctx = talloc_zero(state, struct sdap_deref_ctx); + if (!state->derefctx) goto immediate; + + ret = sysdb_attrs_get_string(group, SYSDB_ORIG_DN, + &state->derefctx->orig_dn); + if (ret != EOK) goto immediate; + + ret = sdap_nested_group_process_deref_step(req); + if (ret != EAGAIN) goto immediate; + } else { + ret = sdap_nested_group_process_step(req); + if (ret != EAGAIN) goto immediate; + } return req; @@ -3418,6 +3456,136 @@ static void sdap_nested_group_process_user(struct tevent_req *subreq); static errno_t sdap_nested_group_lookup_user(struct tevent_req *req, tevent_req_fn fn); static errno_t sdap_nested_group_lookup_group(struct tevent_req *req); +static errno_t sdap_nested_group_process_deref_call(struct tevent_req *req); +static errno_t sdap_nested_group_process_noderef(struct tevent_req *req); + +static errno_t sdap_nested_group_process_deref_step(struct tevent_req *req) +{ + errno_t ret; + struct sdap_nested_group_ctx *state = + tevent_req_data(req, struct sdap_nested_group_ctx); + size_t missing = 0; + struct ldb_message **msgs = NULL; + enum sysdb_member_type mtype; + struct sdap_deref_ctx *dctx = state->derefctx; + + dctx->deref_threshold = dp_opt_get_int(state->opts->basic, + SDAP_DEREF_THRESHOLD); + + dctx->expired_users = talloc_array(dctx, char *, + state->members->num_values + 1); + dctx->expired_groups = talloc_array(dctx, char *, + state->members->num_values + 1); + dctx->missing_dns = talloc_array(dctx, char *, + state->members->num_values + 1); + if (!dctx->expired_users || + !dctx->expired_groups || + !dctx->missing_dns) return ENOMEM; + + while (true) { + if (state->member_index >= state->members->num_values) { + /* No more entries to check. Return success */ + talloc_zfree(state->member_dn); + ret = EOK; + break; + } + + /* Continue to loop through until all entries have been + * processed. + */ + ret = sdap_nested_group_check_hash(state); + if (ret == EOK) { + talloc_zfree(state->member_dn); + break; /* All remaining members in hash, check missing */ + } else if (ret != ENOENT) { + goto done; /* Unexpected error */ + } + + ret = sdap_nested_group_check_cache(state, state->sysdb, + state->domain, state->opts, + state->member_dn, + &msgs, &mtype); + if (ret == EOK) { + /* The entry is cached and valid */ + state->member_index++; + talloc_zfree(state->member_dn); + continue; + } else if (ret == EAGAIN) { + /* The entry is cached but needs refresh */ + switch(mtype) { + case SYSDB_MEMBER_GROUP: + DEBUG(8, ("Cached LDAP group [%s] needs refresh\n", + state->member_dn)); + + missing++; + + dctx->expired_groups[dctx->expired_groups_num] = + talloc_move(dctx, &state->member_dn); + dctx->expired_groups_num++; + + state->member_index++; + continue; + case SYSDB_MEMBER_USER: + DEBUG(8, ("Cached LDAP user [%s] needs refresh\n", + state->member_dn)); + missing++; + + dctx->expired_users[dctx->expired_users_num] = + talloc_move(dctx, &state->member_dn); + dctx->expired_users_num++; + + state->member_index++; + continue; + default: + DEBUG(2, ("Unknown member value\n")); + ret = EINVAL; + goto done; + } + } else if (ret == ENOENT) { + /* The entry is missing. It is unclear whether it + * is a user or a group so we'll need to try looking + * it up */ + missing++; + + dctx->missing_dns[dctx->missing_dns_num] = + talloc_move(dctx, &state->member_dn); + dctx->missing_dns_num++; + + state->member_index++; + continue; + } + + /* Unexpected error, skip this entry */ + state->member_index++; + continue; + } /* while (true) */ + + + dctx->expired_users[dctx->expired_users_num] = NULL; + dctx->expired_groups[dctx->expired_groups_num] = NULL; + dctx->missing_dns[dctx->missing_dns_num] = NULL; + + if (missing == 0) { + ret = EOK; + goto done; + } + + if (missing > dctx->deref_threshold) { + DEBUG(6, ("Missing data past threshold, doing a full deref\n")); + ret = sdap_nested_group_process_deref_call(req); + } else { + DEBUG(6, ("Falling back to individual lookups\n")); + ret = sdap_nested_group_process_noderef(req); + } + + if (ret != EOK && ret != EAGAIN) goto done; + return EAGAIN; + +done: + talloc_zfree(state->member_dn); + return ret; +} + static errno_t sdap_nested_group_process_step(struct tevent_req *req) { @@ -3658,6 +3826,110 @@ done: return ret; } +static void sdap_nested_group_process_deref(struct tevent_req *subreq); + +static errno_t +sdap_nested_group_process_deref_call(struct tevent_req *req) +{ + struct tevent_req *subreq; + struct sdap_attr_map_info *maps; + const char **sdap_attrs; + int ret; + int timeout; + const int num_maps = 2; + struct sdap_nested_group_ctx *state = + tevent_req_data(req, struct sdap_nested_group_ctx); + + maps = talloc_array(state, struct sdap_attr_map_info, num_maps+1); + if (!maps) return ENOMEM; + + maps[0].map = state->opts->user_map; + maps[0].num_attrs = SDAP_OPTS_USER; + maps[1].map = state->opts->group_map; + maps[1].num_attrs = SDAP_OPTS_GROUP; + maps[2].map = NULL; + + /* Pull down the whole group map, but only pull down username + * and originalDN for users. */ + ret = build_attrs_from_map(state, state->opts->group_map, + SDAP_OPTS_GROUP, &sdap_attrs); + if (ret != EOK) goto fail; + + sdap_attrs = talloc_realloc(NULL, sdap_attrs, const char *, + SDAP_OPTS_GROUP + 2); + if (!sdap_attrs) { + ret = ENOMEM; + goto fail; + } + + sdap_attrs[SDAP_OPTS_GROUP] = \ + state->opts->user_map[SDAP_AT_USER_NAME].name; + sdap_attrs[SDAP_OPTS_GROUP + 1] = NULL; + + timeout = dp_opt_get_int(state->opts->basic, SDAP_ENTRY_CACHE_TIMEOUT); + + subreq = sdap_deref_search_send(state, state->ev, state->opts, + state->sh, state->derefctx->orig_dn, + state->opts->group_map[SDAP_AT_GROUP_MEMBER].name, + sdap_attrs, num_maps, maps, timeout); + if (!subreq) { + ret = EIO; + goto fail; + } + talloc_steal(subreq, sdap_attrs); + talloc_steal(subreq, maps); + + tevent_req_set_callback(subreq, sdap_nested_group_process_deref, req); + return EOK; + +fail: + talloc_free(sdap_attrs); + talloc_free(maps); + return ret; +} + +static errno_t sdap_nested_group_process_noderef(struct tevent_req *req) +{ + struct sdap_nested_group_ctx *state = + tevent_req_data(req, struct sdap_nested_group_ctx); + struct sdap_deref_ctx *dctx = state->derefctx; + errno_t ret; + + if (dctx->expired_users_index < dctx->expired_users_num) { + state->member_dn = dctx->expired_users[dctx->expired_users_index]; + DEBUG(8, ("Refreshing expired user [%s]\n", state->member_dn)); + + ret = sdap_nested_group_lookup_user( + req, sdap_nested_group_process_user); + if (ret != EOK) goto done; + return EAGAIN; + } + + if (dctx->expired_groups_index < dctx->expired_groups_num) { + state->member_dn = dctx->expired_groups[dctx->expired_groups_index]; + DEBUG(8, ("Refreshing expired group [%s]\n", state->member_dn)); + + ret = sdap_nested_group_lookup_group(req); + if (ret != EOK) goto done; + return EAGAIN; + } + + if (dctx->missing_dns_index < dctx->missing_dns_num) { + state->member_dn = dctx->missing_dns[dctx->missing_dns_index]; + DEBUG(8, ("Looking up missing DN [%s]\n", state->member_dn)); + + /* Try users first for generic missing DNs */ + ret = sdap_nested_group_lookup_user( + req, sdap_nested_group_process_ldap_user); + if (ret != EOK) goto done; + return EAGAIN; + } + + ret = EOK; +done: + return ret; +} + static errno_t sdap_nested_group_lookup_user(struct tevent_req *req, tevent_req_fn fn) { @@ -3803,9 +4075,15 @@ static void sdap_nested_group_process_user(struct tevent_req *subreq) talloc_steal(state->users, replies[0]); skip: - state->member_index++; - talloc_zfree(state->member_dn); - ret = sdap_nested_group_process_step(req); + if (state->derefctx) { + state->derefctx->expired_users_index++; + ret = sdap_nested_group_process_noderef(req); + } else { + state->member_index++; + talloc_zfree(state->member_dn); + ret = sdap_nested_group_process_step(req); + } + if (ret == EOK) { /* EOK means it's complete */ tevent_req_done(req); @@ -3875,9 +4153,15 @@ static void sdap_nested_group_process_group(struct tevent_req *subreq) return; skip: - state->member_index++; - talloc_zfree(state->member_dn); - ret = sdap_nested_group_process_step(req); + if (state->derefctx) { + state->derefctx->expired_groups_index++; + ret = sdap_nested_group_process_noderef(req); + } else { + state->member_index++; + talloc_zfree(state->member_dn); + ret = sdap_nested_group_process_step(req); + } + if (ret == EOK) { /* EOK means it's complete */ tevent_req_done(req); @@ -3908,9 +4192,22 @@ static void sdap_group_internal_nesting_done(struct tevent_req *subreq) return; } - state->member_index++; - talloc_zfree(state->member_dn); - ret = sdap_nested_group_process_step(req); + if (state->derefctx) { + if (state->derefctx->expired_groups_index < + state->derefctx->expired_groups_num) { + state->derefctx->expired_groups_index++; + } else { + state->derefctx->missing_dns_index++; + } + + state->derefctx->expired_users_index++; + ret = sdap_nested_group_process_noderef(req); + } else { + state->member_index++; + talloc_zfree(state->member_dn); + ret = sdap_nested_group_process_step(req); + } + if (ret == EOK) { /* EOK means it's complete */ tevent_req_done(req); @@ -3984,9 +4281,15 @@ static void sdap_nested_group_process_ldap_user(struct tevent_req *subreq) talloc_steal(state->users, replies[0]); /* Move on to the next member */ - state->member_index++; - talloc_zfree(state->member_dn); - ret = sdap_nested_group_process_step(req); + if (state->derefctx) { + state->derefctx->missing_dns_index++; + ret = sdap_nested_group_process_noderef(req); + } else { + state->member_index++; + talloc_zfree(state->member_dn); + ret = sdap_nested_group_process_step(req); + } + if (ret == EOK) { /* EOK means it's complete */ tevent_req_done(req); @@ -4002,6 +4305,143 @@ done: talloc_free(tmp_ctx); } +static errno_t +sdap_nested_group_process_deref_result(struct tevent_req *req); + +static void sdap_nested_group_process_deref(struct tevent_req *subreq) +{ + errno_t ret; + struct tevent_req *req = + tevent_req_callback_data(subreq, struct tevent_req); + struct sdap_nested_group_ctx *state = + tevent_req_data(req, struct sdap_nested_group_ctx); + + ret = sdap_deref_search_recv(subreq, state->derefctx, + &state->derefctx->num_results, + &state->derefctx->deref_result); + talloc_zfree(subreq); + if (ret != EOK && ret != ENOENT) { + tevent_req_error(req, ret); + return; + } else if (ret == ENOENT || state->derefctx->deref_result == 0) { + /* Nothing could be dereferenced. Done. */ + tevent_req_done(req); + return; + } + + state->derefctx->result_index = 0; + + DEBUG(8, ("Received %d dereference results, about to process them\n", + state->derefctx->num_results)); + ret = sdap_nested_group_process_deref_result(req); + if (ret == EOK) { + tevent_req_done(req); + } else if (ret != EAGAIN) { + tevent_req_error(req, ret); + } + + /* EAGAIN means a recursive search is in progress */ +} + +static void +sdap_nested_group_process_deref_recurse_done(struct tevent_req *subreq); + +static errno_t +sdap_nested_group_process_deref_result(struct tevent_req *req) +{ + struct sdap_nested_group_ctx *state = + tevent_req_data(req, struct sdap_nested_group_ctx); + struct tevent_req *subreq; + hash_key_t key; + hash_value_t value; + int hret; + const char *orig_dn; + errno_t ret; + struct sdap_deref_ctx *dctx = state->derefctx; + + while (dctx->result_index < dctx->num_results) { + if (dctx->deref_result[dctx->result_index]->map == \ + state->opts->user_map) { + /* Add to appropriate hash table */ + ret = sysdb_attrs_get_string( + dctx->deref_result[dctx->result_index]->attrs, + SYSDB_ORIG_DN, &orig_dn); + if (ret != EOK) { + DEBUG(2, ("The entry has no originalDN\n")); + return ret; + } + + DEBUG(9, ("Found member user [%s]\n", orig_dn)); + + key.type = HASH_KEY_STRING; + key.str = talloc_strdup(state, orig_dn); + + value.type = HASH_VALUE_PTR; + value.ptr = dctx->deref_result[dctx->result_index]->attrs; + + hret = hash_enter(state->users, &key, &value); + if (hret != HASH_SUCCESS) return EIO; + + talloc_steal(state->users, + dctx->deref_result[dctx->result_index]->attrs); + dctx->result_index++; + } else if (dctx->deref_result[dctx->result_index]->map == \ + state->opts->group_map) { + DEBUG(6, ("Recursing down a nested group\n")); + subreq = sdap_nested_group_process_send(state, state->ev, + state->domain, state->sysdb, + dctx->deref_result[dctx->result_index]->attrs, + state->users, state->groups, + state->opts, state->sh, + state->nesting_level + 1); + if (!subreq) return EIO; + + tevent_req_set_callback(subreq, + sdap_nested_group_process_deref_recurse_done, + req); + return EAGAIN; + } else { + /* This should never happen, but if it does, + * do not loop forever */ + DEBUG(2, ("Entry does not match any known map, skipping\n")); + dctx->result_index++; + continue; + } + } + + /* All deref results processed */ + DEBUG(8, ("All dereference results processed\n")); + return EOK; +} + +static void +sdap_nested_group_process_deref_recurse_done(struct tevent_req *subreq) +{ + errno_t ret; + struct tevent_req *req = + tevent_req_callback_data(subreq, struct tevent_req); + struct sdap_nested_group_ctx *state = + tevent_req_data(req, struct sdap_nested_group_ctx); + + ret = sdap_nested_group_process_recv(subreq); + talloc_zfree(subreq); + if (ret != EOK) { + tevent_req_error(req, ret); + return; + } + + state->derefctx->result_index++; + + ret = sdap_nested_group_process_deref_result(req); + if (ret == EOK) { + tevent_req_done(req); + } else if (ret != EAGAIN) { + tevent_req_error(req, ret); + } + + /* EAGAIN means a recursive search is in progress */ +} + static errno_t sdap_nested_group_process_recv(struct tevent_req *req) { TEVENT_REQ_RETURN_ON_ERROR(req); |