summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorStephen Gallagher <sgallagh@redhat.com>2010-08-27 09:13:59 -0400
committerStephen Gallagher <sgallagh@redhat.com>2010-09-08 13:30:33 -0400
commitc53ed27b33ecc7fcce62d4b3a3e55ce9cda1ca7c (patch)
tree959ac4701e208d5d2cc04f4811f7d6253762ce55
parent8443d24c0584f45151e0c80506d7a572b8a38ed7 (diff)
downloadsssd-c53ed27b33ecc7fcce62d4b3a3e55ce9cda1ca7c.tar.gz
sssd-c53ed27b33ecc7fcce62d4b3a3e55ce9cda1ca7c.tar.xz
sssd-c53ed27b33ecc7fcce62d4b3a3e55ce9cda1ca7c.zip
Handle multiple simultaneous enumeration requests
Previously, if a second enumeration request arrived while one was already being processed, each process would receive only a subset of the total number of available users or groups. This is because we were maintaining the response object as a global value in the NSS responder. The second request would come in, see that the data set was already populated, and start reading from wherever the cursor was currently pointed. With this patch, we now move the cursor to the client context instead of the global NSS context. Additionally, this patch completely rewrites the approach to enumerations in the tevent_req style. This makes it much easier to follow in the code. In order to ensure that a slow or malicious client cannot hold onto a reference for the setent result object indefinitely, we set an expiration on the object. We use the enum_cache_timeout here, since that is an appropriate value. If the timeout fires during the normal operation of the get*ent() loop of a client program, we will save the current values of the read index so that we can resume as soon as the object has been refreshed by an implicit setent call. Instead of deleting the enumeration result object immediately after the last in-progress client has read it, we'll keep the object around for the lifetime of enum_cache_timeout. This way, additional clients making enumeration requests can still access the results in-memory.
-rw-r--r--src/responder/common/responder.h11
-rw-r--r--src/responder/nss/nsssrv_cmd.c995
2 files changed, 717 insertions, 289 deletions
diff --git a/src/responder/common/responder.h b/src/responder/common/responder.h
index deb1e5a3c..0f59ffd43 100644
--- a/src/responder/common/responder.h
+++ b/src/responder/common/responder.h
@@ -92,6 +92,11 @@ struct resp_ctx {
void *pvt_ctx;
};
+/* Needed for the NSS responder */
+struct getent_ref_tracker {
+ void *pvt;
+};
+
struct cli_ctx {
struct tevent_context *ev;
struct resp_ctx *rctx;
@@ -104,6 +109,12 @@ struct cli_ctx {
int32_t client_euid;
int32_t client_egid;
int32_t client_pid;
+
+ int pwent_dom_idx;
+ int pwent_cur;
+
+ int grent_dom_idx;
+ int grent_cur;
};
struct sss_cmd_table {
diff --git a/src/responder/nss/nsssrv_cmd.c b/src/responder/nss/nsssrv_cmd.c
index 719608c0c..6df705fb6 100644
--- a/src/responder/nss/nsssrv_cmd.c
+++ b/src/responder/nss/nsssrv_cmd.c
@@ -35,18 +35,20 @@ struct nss_cmd_ctx {
bool check_next;
bool enum_cached;
+ int saved_dom_idx;
+ int saved_cur;
};
struct dom_ctx {
struct sss_domain_info *domain;
struct ldb_result *res;
- int cur;
};
struct getent_ctx {
struct dom_ctx *doms;
int num;
- int cur;
+ bool ready;
+ struct setent_req_list *reqs;
};
struct nss_dom_ctx {
@@ -172,6 +174,63 @@ static int nss_cmd_done(struct nss_cmd_ctx *cmdctx, int ret)
return EOK;
}
+/***************************
+ * Enumeration procedures *
+ ***************************/
+
+struct setent_req_list {
+ struct setent_req_list *prev;
+ struct setent_req_list *next;
+ struct getent_ctx *getent_ctx;
+
+ struct tevent_req *req;
+};
+
+static int
+setent_remove_ref(TALLOC_CTX *ctx);
+static errno_t
+setent_add_ref(TALLOC_CTX *memctx,
+ struct getent_ctx *getent_ctx,
+ struct tevent_req *req)
+{
+ struct setent_req_list *entry =
+ talloc_zero(memctx, struct setent_req_list);
+ if (!entry) {
+ return ENOMEM;
+ }
+
+ entry->req = req;
+ entry->getent_ctx = getent_ctx;
+ DLIST_ADD_END(getent_ctx->reqs, entry, struct setent_req_list *);
+
+ talloc_set_destructor((TALLOC_CTX *)entry, setent_remove_ref);
+ return EOK;
+}
+
+static int
+setent_remove_ref(TALLOC_CTX *ctx)
+{
+ struct setent_req_list *entry =
+ talloc_get_type(ctx, struct setent_req_list);
+ DLIST_REMOVE(entry->getent_ctx->reqs, entry);
+ return 0;
+}
+
+struct setent_ctx {
+ struct cli_ctx *client;
+ struct nss_ctx *nctx;
+ struct nss_dom_ctx *dctx;
+ struct getent_ctx *getent_ctx;
+};
+
+struct setent_step_ctx {
+ struct nss_ctx *nctx;
+ struct nss_dom_ctx *dctx;
+ struct getent_ctx *getent_ctx;
+ struct resp_ctx *rctx;
+ bool enum_cached;
+};
+
/****************************************************************************
* PASSWD db related functions
***************************************************************************/
@@ -944,32 +1003,176 @@ done:
* the last returned user.
*/
static int nss_cmd_getpwent_immediate(struct nss_cmd_ctx *cmdctx);
+struct tevent_req * nss_cmd_setpwent_send(TALLOC_CTX *mem_ctx,
+ struct cli_ctx *client);
+static void nss_cmd_setpwent_done(struct tevent_req *req);
+static int nss_cmd_setpwent(struct cli_ctx *cctx)
+{
+ struct nss_cmd_ctx *cmdctx;
+ struct tevent_req *req;
+ errno_t ret = EOK;
-static void nss_cmd_getpwent_dp_callback(uint16_t err_maj, uint32_t err_min,
- const char *err_msg, void *ptr);
+ cmdctx = talloc_zero(cctx, struct nss_cmd_ctx);
+ if (!cmdctx) {
+ return ENOMEM;
+ }
+ cmdctx->cctx = cctx;
+
+ req = nss_cmd_setpwent_send(cmdctx, cctx);
+ if (!req) {
+ DEBUG(0, ("Fatal error calling nss_cmd_setpwent_send\n"));
+ ret = EIO;
+ goto done;
+ }
+ tevent_req_set_callback(req, nss_cmd_setpwent_done, cmdctx);
+
+done:
+ return nss_cmd_done(cmdctx, ret);
+}
-static int nss_cmd_getpwent_search(struct nss_dom_ctx *dctx)
+static errno_t nss_cmd_setpwent_step(struct setent_step_ctx *step_ctx);
+struct tevent_req *nss_cmd_setpwent_send(TALLOC_CTX *mem_ctx,
+ struct cli_ctx *client)
{
- struct nss_cmd_ctx *cmdctx = dctx->cmdctx;
- struct sss_domain_info *dom = dctx->domain;
- struct cli_ctx *cctx = cmdctx->cctx;
- struct sysdb_ctx *sysdb;
- struct ldb_result *res;
- struct getent_ctx *pctx;
+ errno_t ret;
struct nss_ctx *nctx;
- int timeout;
- int ret;
+ struct tevent_req *req;
+ struct setent_ctx *state;
+ struct sss_domain_info *dom;
+ struct setent_step_ctx *step_ctx;
+ bool enum_cached = false;
+ time_t now = time(NULL);
- nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
- pctx = nctx->pctx;
- if (pctx == NULL) {
- pctx = talloc_zero(nctx, struct getent_ctx);
- if (!pctx) {
- return ENOMEM;
+ DEBUG(4, ("Received setpwent request\n"));
+ nctx = talloc_get_type(client->rctx->pvt_ctx, struct nss_ctx);
+
+ /* Reset the read pointers */
+ client->pwent_dom_idx = 0;
+ client->pwent_cur = 0;
+
+ req = tevent_req_create(mem_ctx, &state, struct setent_ctx);
+ if (!req) {
+ DEBUG(0, ("Could not create tevent request for setpwent\n"));
+ return NULL;
+ }
+
+ state->nctx = nctx;
+ state->client = client;
+
+ /* do not query backends if we have a recent enumeration */
+ if (nctx->enum_cache_timeout) {
+ if (nctx->last_user_enum + nctx->enum_cache_timeout > now) {
+ enum_cached = true;
+ }
+ }
+
+ state->dctx = talloc_zero(state, struct nss_dom_ctx);
+ if (!state->dctx) {
+ ret = ENOMEM;
+ goto error;
+ }
+
+ /* check if enumeration is enabled in any domain */
+ for (dom = client->rctx->domains; dom; dom = dom->next) {
+ if (dom->enumerate != 0) break;
+ }
+ state->dctx->domain = dom;
+
+ if (state->dctx->domain == NULL) {
+ DEBUG(2, ("Enumeration disabled on all domains!\n"));
+ ret = ENOENT;
+ goto error;
+ }
+
+ state->dctx->check_provider =
+ NEED_CHECK_PROVIDER(state->dctx->domain->provider);
+
+ /* Is the result context already available */
+ if (state->nctx->pctx) {
+ if (state->nctx->pctx->ready) {
+ /* All of the necessary data is in place
+ * We can return now, getpwent requests will work at this point
+ */
+ tevent_req_done(req);
+ tevent_req_post(req, state->nctx->rctx->ev);
+ }
+ else {
+ /* Object is still being constructed
+ * Register for notification when it's
+ * ready.
+ */
+ ret = setent_add_ref(state->client, state->nctx->pctx, req);
+ if (ret != EOK) {
+ talloc_free(req);
+ return NULL;
+ }
}
- nctx->pctx = pctx;
+ return req;
}
+ /* Create a new result context
+ * We are creating it on the nss_ctx so that it doesn't
+ * go away if the original request does. We will delete
+ * it when the refcount goes to zero;
+ */
+ state->nctx->pctx = talloc_zero(nctx, struct getent_ctx);
+ if (!state->nctx->pctx) {
+ ret = ENOMEM;
+ goto error;
+ }
+ state->getent_ctx = nctx->pctx;
+
+ /* Add a callback reference for ourselves */
+ setent_add_ref(state->client, state->nctx->pctx, req);
+
+ /* ok, start the searches */
+ step_ctx = talloc_zero(state->getent_ctx, struct setent_step_ctx);
+ if (!step_ctx) {
+ ret = ENOMEM;
+ goto error;
+ }
+
+ /* Steal the dom_ctx onto the step_ctx so it doesn't go out of scope if
+ * this request is canceled while other requests are in-progress.
+ */
+ step_ctx->dctx = talloc_steal(step_ctx, state->dctx);
+ step_ctx->nctx = state->nctx;
+ step_ctx->getent_ctx = state->getent_ctx;
+ step_ctx->rctx = client->rctx;
+ step_ctx->enum_cached = enum_cached;
+
+ ret = nss_cmd_setpwent_step(step_ctx);
+ if (ret != EOK) goto error;
+
+ return req;
+
+ error:
+ tevent_req_error(req, ret);
+ tevent_req_post(req, client->rctx->ev);
+ return req;
+}
+
+static void nss_cmd_setpwent_dp_callback(uint16_t err_maj, uint32_t err_min,
+ const char *err_msg, void *ptr);
+static void setpwent_result_timeout(struct tevent_context *ev,
+ struct tevent_timer *te,
+ struct timeval current_time,
+ void *pvt);
+static errno_t nss_cmd_setpwent_step(struct setent_step_ctx *step_ctx)
+{
+ errno_t ret;
+ struct sss_domain_info *dom = step_ctx->dctx->domain;
+ struct resp_ctx *rctx = step_ctx->rctx;
+ struct nss_dom_ctx *dctx = step_ctx->dctx;
+ struct getent_ctx *pctx = step_ctx->getent_ctx;
+ struct nss_ctx *nctx = step_ctx->nctx;
+ struct sysdb_ctx *sysdb;
+ struct ldb_result *res;
+ struct setent_req_list *req;
+ struct timeval tv;
+ struct tevent_timer *te;
+ int timeout;
+
while (dom) {
while (dom && dom->enumerate == 0) {
dom = dom->next;
@@ -980,7 +1183,7 @@ static int nss_cmd_getpwent_search(struct nss_dom_ctx *dctx)
if (dom != dctx->domain) {
/* make sure we reset the check_provider flag when we check
* a new domain */
- if (cmdctx->enum_cached) {
+ if (step_ctx->enum_cached) {
dctx->check_provider = false;
} else {
dctx->check_provider = NEED_CHECK_PROVIDER(dom->provider);
@@ -990,9 +1193,9 @@ static int nss_cmd_getpwent_search(struct nss_dom_ctx *dctx)
/* make sure to update the dctx if we changed domain */
dctx->domain = dom;
- DEBUG(4, ("Requesting info for domain [%s]\n", dom->name));
+ DEBUG(6, ("Requesting info for domain [%s]\n", dom->name));
- ret = sysdb_get_ctx_from_list(cctx->rctx->db_list, dom, &sysdb);
+ ret = sysdb_get_ctx_from_list(rctx->db_list, dom, &sysdb);
if (ret != EOK) {
DEBUG(0, ("Fatal: Sysdb CTX not found for this domain!\n"));
return EIO;
@@ -1001,10 +1204,12 @@ static int nss_cmd_getpwent_search(struct nss_dom_ctx *dctx)
/* if this is a caching provider (or if we haven't checked the cache
* yet) then verify that the cache is uptodate */
if (dctx->check_provider) {
+ /* Only do this once per provider */
dctx->check_provider = false;
timeout = SSS_CLI_SOCKET_TIMEOUT;
- ret = sss_dp_send_acct_req(cctx->rctx, cmdctx,
- nss_cmd_getpwent_dp_callback, dctx,
+
+ ret = sss_dp_send_acct_req(rctx, step_ctx,
+ nss_cmd_setpwent_dp_callback, step_ctx,
timeout, dom->name, true,
SSS_DP_USER, NULL, 0);
if (ret == EOK) {
@@ -1030,7 +1235,7 @@ static int nss_cmd_getpwent_search(struct nss_dom_ctx *dctx)
continue;
}
- pctx->doms = talloc_realloc(pctx, pctx->doms,
+ nctx->pctx->doms = talloc_realloc(pctx, pctx->doms,
struct dom_ctx, pctx->num +1);
if (!pctx->doms) {
talloc_free(pctx);
@@ -1038,41 +1243,65 @@ static int nss_cmd_getpwent_search(struct nss_dom_ctx *dctx)
return ENOMEM;
}
- pctx->doms[pctx->num].domain = dctx->domain;
- pctx->doms[pctx->num].res = talloc_steal(pctx->doms, res);
- pctx->doms[pctx->num].cur = 0;
+ nctx->pctx->doms[pctx->num].domain = dctx->domain;
+ nctx->pctx->doms[pctx->num].res = talloc_steal(pctx->doms, res);
- pctx->num++;
+ nctx->pctx->num++;
/* do not reply until all domain searches are done */
dom = dom->next;
}
- /* set cache mark */
- nctx->last_user_enum = time(NULL);
+ /* We've finished all our lookups
+ * The result object is now safe to read.
+ */
+ nctx->pctx->ready = true;
- if (cmdctx->immediate) {
- /* this was a getpwent call w/o setpwent,
- * return immediately one result */
- return nss_cmd_getpwent_immediate(cmdctx);
+ /* Set up a lifetime timer for this result object
+ * We don't want this result object to outlive the
+ * enum cache refresh timeout
+ */
+ tv = tevent_timeval_current_ofs(nctx->enum_cache_timeout, 0);
+ te = tevent_add_timer(rctx->ev, nctx->pctx, tv,
+ setpwent_result_timeout, nctx);
+ if (!te) {
+ DEBUG(0, ("Could not set up life timer for setpwent result object. "
+ "Entries may become stale.\n"));
}
- /* create response packet */
- ret = sss_packet_new(cctx->creq, 0,
- sss_packet_get_cmd(cctx->creq->in),
- &cctx->creq->out);
- if (ret == EOK) {
- sss_cmd_done(cctx, cmdctx);
+ /* Notify the waiting clients */
+ while (nctx->pctx->reqs) {
+ tevent_req_done(nctx->pctx->reqs->req);
+ /* Freeing each entry in the list removes it from the dlist */
+ req = nctx->pctx->reqs;
+ nctx->pctx->reqs = nctx->pctx->reqs->next;
+ talloc_free(req);
}
- return ret;
+ return EOK;
}
-static void nss_cmd_getpwent_dp_callback(uint16_t err_maj, uint32_t err_min,
+static void setpwent_result_timeout(struct tevent_context *ev,
+ struct tevent_timer *te,
+ struct timeval current_time,
+ void *pvt)
+{
+ struct nss_ctx *nctx = talloc_get_type(pvt, struct nss_ctx);
+
+ DEBUG(1, ("setpwent result object has expired. Cleaning up.\n"));
+
+ /* Free the passwd enumeration context.
+ * If additional getpwent requests come in, they will invoke
+ * an implicit setpwent and refresh the result object.
+ */
+ talloc_zfree(nctx->pctx);
+}
+
+static void nss_cmd_setpwent_dp_callback(uint16_t err_maj, uint32_t err_min,
const char *err_msg, void *ptr)
{
- struct nss_dom_ctx *dctx = talloc_get_type(ptr, struct nss_dom_ctx);
- struct nss_cmd_ctx *cmdctx = dctx->cmdctx;
- struct cli_ctx *cctx = cmdctx->cctx;
+ struct setent_step_ctx *step_ctx =
+ talloc_get_type(ptr, struct setent_step_ctx);
+ struct getent_ctx *pctx = step_ctx->nctx->pctx;
int ret;
if (err_maj) {
@@ -1082,84 +1311,117 @@ static void nss_cmd_getpwent_dp_callback(uint16_t err_maj, uint32_t err_min,
(unsigned int)err_maj, (unsigned int)err_min, err_msg));
}
- ret = nss_cmd_getpwent_search(dctx);
-
+ ret = nss_cmd_setpwent_step(step_ctx);
if (ret) {
- NSS_CMD_FATAL_ERROR(cctx);
+ /* Notify any waiting processes of failure */
+ while(step_ctx->nctx->pctx->reqs) {
+ tevent_req_error(pctx->reqs->req, ret);
+ /* Freeing each entry in the list removes it from the dlist */
+ talloc_free(pctx->reqs);
+ }
}
}
-static int nss_cmd_setpwent_ext(struct cli_ctx *cctx, bool immediate)
+static errno_t nss_cmd_setpwent_recv(struct tevent_req *req)
{
- struct sss_domain_info *dom;
- struct nss_cmd_ctx *cmdctx;
- struct nss_dom_ctx *dctx;
- struct nss_ctx *nctx;
- time_t now = time(NULL);
- int ret;
+ TEVENT_REQ_RETURN_ON_ERROR(req);
+ return EOK;
+}
- DEBUG(4, ("Requesting info for all users\n"));
+static void nss_cmd_setpwent_done(struct tevent_req *req)
+{
+ errno_t ret;
+ struct nss_cmd_ctx *cmdctx =
+ tevent_req_callback_data(req, struct nss_cmd_ctx);
- nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
- talloc_free(nctx->pctx);
- nctx->pctx = NULL;
+ ret = nss_cmd_setpwent_recv(req);
+ talloc_zfree(req);
+ if (ret == EOK || ret == ENOENT) {
+ /* Either we succeeded or no domains were eligible */
+ ret = sss_packet_new(cmdctx->cctx->creq, 0,
+ sss_packet_get_cmd(cmdctx->cctx->creq->in),
+ &cmdctx->cctx->creq->out);
+ if (ret == EOK) {
+ sss_cmd_done(cmdctx->cctx, cmdctx);
+ return;
+ }
+ }
- cmdctx = talloc_zero(cctx, struct nss_cmd_ctx);
+ /* Something bad happened */
+ nss_cmd_done(cmdctx, ret);
+}
+
+static void nss_cmd_implicit_setpwent_done(struct tevent_req *req);
+static int nss_cmd_getpwent(struct cli_ctx *cctx)
+{
+ struct nss_ctx *nctx;
+ struct nss_cmd_ctx *cmdctx;
+ struct tevent_req *req;
+
+ DEBUG(4, ("Requesting info for all accounts\n"));
+
+ cmdctx = talloc(cctx, struct nss_cmd_ctx);
if (!cmdctx) {
return ENOMEM;
}
cmdctx->cctx = cctx;
- cmdctx->immediate = immediate;
- dctx = talloc_zero(cmdctx, struct nss_dom_ctx);
- if (!dctx) {
- ret = ENOMEM;
- goto done;
- }
- dctx->cmdctx = cmdctx;
+ /* Save the current index and cursor locations
+ * If we end up calling setpwent implicitly, because the response object
+ * expired and has to be recreated, we want to resume from the same
+ * location.
+ */
+ cmdctx->saved_dom_idx = cctx->pwent_dom_idx;
+ cmdctx->saved_cur = cctx->pwent_cur;
- /* do not query backends if we have a recent enumeration */
- if (nctx->enum_cache_timeout) {
- if (nctx->last_user_enum + nctx->enum_cache_timeout > now) {
- cmdctx->enum_cached = true;
+ nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
+ if(!nctx->pctx || !nctx->pctx->ready) {
+ /* Make sure we invoke setpwent if it hasn't been run or is still
+ * processing from another client
+ */
+ req = nss_cmd_setpwent_send(cctx, cctx);
+ if (!req) {
+ return EIO;
}
+ tevent_req_set_callback(req, nss_cmd_implicit_setpwent_done, cmdctx);
+ return EOK;
}
- /* check if enumeration is enabled in any domain */
- for (dom = cctx->rctx->domains; dom; dom = dom->next) {
- if (dom->enumerate != 0) break;
+ return nss_cmd_getpwent_immediate(cmdctx);
+}
+
+static int nss_cmd_retpwent(struct cli_ctx *cctx, int num);
+static int nss_cmd_getpwent_immediate(struct nss_cmd_ctx *cmdctx)
+{
+ struct cli_ctx *cctx = cmdctx->cctx;
+ uint8_t *body;
+ size_t blen;
+ uint32_t num;
+ int ret;
+
+ /* get max num of entries to return in one call */
+ sss_packet_get_body(cctx->creq->in, &body, &blen);
+ if (blen != sizeof(uint32_t)) {
+ return EINVAL;
}
- dctx->domain = dom;
+ num = *((uint32_t *)body);
- if (dctx->domain == NULL) {
- DEBUG(2, ("Enumeration disabled on all domains!\n"));
- if (cmdctx->immediate) {
- ret = ENOENT;
- } else {
- ret = sss_packet_new(cctx->creq, 0,
- sss_packet_get_cmd(cctx->creq->in),
- &cctx->creq->out);
- if (ret == EOK) {
- sss_cmd_done(cctx, cmdctx);
- }
- }
- goto done;
+ /* create response packet */
+ ret = sss_packet_new(cctx->creq, 0,
+ sss_packet_get_cmd(cctx->creq->in),
+ &cctx->creq->out);
+ if (ret != EOK) {
+ return ret;
}
- dctx->check_provider = NEED_CHECK_PROVIDER(dctx->domain->provider);
+ ret = nss_cmd_retpwent(cctx, num);
- /* ok, start the searches */
- ret = nss_cmd_getpwent_search(dctx);
-done:
- return nss_cmd_done(cmdctx, ret);
-}
+ sss_packet_set_error(cctx->creq->out, ret);
+ sss_cmd_done(cctx, cmdctx);
-static int nss_cmd_setpwent(struct cli_ctx *cctx)
-{
- return nss_cmd_setpwent_ext(cctx, false);
+ return EOK;
}
-
static int nss_cmd_retpwent(struct cli_ctx *cctx, int num)
{
struct nss_ctx *nctx;
@@ -1175,26 +1437,26 @@ static int nss_cmd_retpwent(struct cli_ctx *cctx, int num)
pctx = nctx->pctx;
while (ret == ENOENT) {
- if (pctx->cur >= pctx->num) break;
+ if (cctx->pwent_dom_idx >= pctx->num) break;
- pdom = &pctx->doms[pctx->cur];
+ pdom = &pctx->doms[cctx->pwent_dom_idx];
- n = pdom->res->count - pdom->cur;
- if (n == 0 && (pctx->cur+1 < pctx->num)) {
- pctx->cur++;
- pdom = &pctx->doms[pctx->cur];
- n = pdom->res->count - pdom->cur;
+ n = pdom->res->count - cctx->pwent_cur;
+ if (n <= 0 && (cctx->pwent_dom_idx+1 < pctx->num)) {
+ cctx->pwent_dom_idx++;
+ pdom = &pctx->doms[cctx->pwent_dom_idx];
+ n = pdom->res->count - cctx->pwent_cur;
}
if (!n) break;
if (n > num) n = num;
- msgs = &(pdom->res->msgs[pdom->cur]);
+ msgs = &(pdom->res->msgs[cctx->pwent_cur]);
ret = fill_pwent(cctx->creq->out, pdom->domain, nctx, true, msgs, &n);
- pdom->cur += n;
+ cctx->pwent_cur += n;
}
none:
@@ -1204,58 +1466,35 @@ none:
return ret;
}
-static int nss_cmd_getpwent_immediate(struct nss_cmd_ctx *cmdctx)
+static void nss_cmd_implicit_setpwent_done(struct tevent_req *req)
{
- struct cli_ctx *cctx = cmdctx->cctx;
- uint8_t *body;
- size_t blen;
- uint32_t num;
- int ret;
+ errno_t ret;
+ struct nss_cmd_ctx *cmdctx =
+ tevent_req_callback_data(req, struct nss_cmd_ctx);
- /* get max num of entries to return in one call */
- sss_packet_get_body(cctx->creq->in, &body, &blen);
- if (blen != sizeof(uint32_t)) {
- return EINVAL;
- }
- num = *((uint32_t *)body);
+ ret = nss_cmd_setpwent_recv(req);
+ talloc_zfree(req);
- /* create response packet */
- ret = sss_packet_new(cctx->creq, 0,
- sss_packet_get_cmd(cctx->creq->in),
- &cctx->creq->out);
- if (ret != EOK) {
- return ret;
+ /* ENOENT is acceptable, as it just means that there were no entries
+ * to be returned. This will be handled gracefully in nss_cmd_retpwent
+ * later.
+ */
+ if (ret != EOK && ret != ENOENT) {
+ DEBUG(0, ("Implicit setpwent failed with unexpected error [%d][%s]\n",
+ ret, strerror(ret)));
+ NSS_CMD_FATAL_ERROR(cmdctx);
}
- ret = nss_cmd_retpwent(cctx, num);
-
- sss_packet_set_error(cctx->creq->out, ret);
- sss_cmd_done(cctx, cmdctx);
+ /* Restore the saved index and cursor locations */
+ cmdctx->cctx->pwent_dom_idx = cmdctx->saved_dom_idx;
+ cmdctx->cctx->pwent_cur = cmdctx->saved_cur;
- return EOK;
-}
-
-static int nss_cmd_getpwent(struct cli_ctx *cctx)
-{
- struct nss_ctx *nctx;
- struct nss_cmd_ctx *cmdctx;
-
- DEBUG(4, ("Requesting info for all accounts\n"));
-
- nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
-
- /* see if we need to trigger an implicit setpwent() */
- if (nctx->pctx == NULL) {
- return nss_cmd_setpwent_ext(cctx, true);
- }
-
- cmdctx = talloc(cctx, struct nss_cmd_ctx);
- if (!cmdctx) {
- return ENOMEM;
+ ret = nss_cmd_getpwent_immediate(cmdctx);
+ if (ret != EOK) {
+ DEBUG(0, ("Immediate retrieval failed with unexpected error "
+ "[%d][%s]\n", ret, strerror(ret)));
+ NSS_CMD_FATAL_ERROR(cmdctx);
}
- cmdctx->cctx = cctx;
-
- return nss_cmd_getpwent_immediate(cmdctx);
}
static int nss_cmd_endpwent(struct cli_ctx *cctx)
@@ -1277,9 +1516,9 @@ static int nss_cmd_endpwent(struct cli_ctx *cctx)
}
if (nctx->pctx == NULL) goto done;
- /* free results and reset */
- talloc_free(nctx->pctx);
- nctx->pctx = NULL;
+ /* Reset the indices so that subsequent requests start at zero */
+ cctx->pwent_dom_idx = 0;
+ cctx->pwent_cur = 0;
done:
sss_cmd_done(cctx, NULL);
@@ -2014,33 +2253,176 @@ done:
* even if the data is still being fetched
* - make getgrent() wait on the mutex
*/
-static int nss_cmd_getgrent_immediate(struct nss_cmd_ctx *cmdctx);
+struct tevent_req *nss_cmd_setgrent_send(TALLOC_CTX *mem_ctx,
+ struct cli_ctx *client);
+static void nss_cmd_setgrent_done(struct tevent_req *req);
+static int nss_cmd_setgrent(struct cli_ctx *cctx)
+{
+ struct nss_cmd_ctx *cmdctx;
+ struct tevent_req *req;
+ errno_t ret = EOK;
-static void nss_cmd_getgrent_dp_callback(uint16_t err_maj, uint32_t err_min,
- const char *err_msg, void *ptr);
+ cmdctx = talloc_zero(cctx, struct nss_cmd_ctx);
+ if (!cmdctx) {
+ return ENOMEM;
+ }
+ cmdctx->cctx = cctx;
+
+ req = nss_cmd_setgrent_send(cmdctx, cctx);
+ if (!req) {
+ DEBUG(0, ("Fatal error calling nss_cmd_setgrent_send\n"));
+ ret = EIO;
+ goto done;
+ }
+ tevent_req_set_callback(req, nss_cmd_setgrent_done, cmdctx);
+
+done:
+ return nss_cmd_done(cmdctx, ret);
+}
-static int nss_cmd_getgrent_search(struct nss_dom_ctx *dctx)
+static errno_t nss_cmd_setgrent_step(struct setent_step_ctx *step_ctx);
+struct tevent_req *nss_cmd_setgrent_send(TALLOC_CTX *mem_ctx,
+ struct cli_ctx *client)
{
- struct nss_cmd_ctx *cmdctx = dctx->cmdctx;
- struct sss_domain_info *dom = dctx->domain;
- struct cli_ctx *cctx = cmdctx->cctx;
- struct sysdb_ctx *sysdb;
- struct ldb_result *res;
- struct getent_ctx *gctx;
+ errno_t ret;
struct nss_ctx *nctx;
- int timeout;
- int ret;
+ struct tevent_req *req;
+ struct setent_ctx *state;
+ struct sss_domain_info *dom;
+ struct setent_step_ctx *step_ctx;
+ bool enum_cached = false;
+ time_t now = time(NULL);
- nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
- gctx = nctx->gctx;
- if (gctx == NULL) {
- gctx = talloc_zero(nctx, struct getent_ctx);
- if (!gctx) {
- return ENOMEM;
+ DEBUG(4, ("Received setgrent request\n"));
+ nctx = talloc_get_type(client->rctx->pvt_ctx, struct nss_ctx);
+
+ /* Reset the read pointers */
+ client->grent_dom_idx = 0;
+ client->grent_cur = 0;
+
+ req = tevent_req_create(mem_ctx, &state, struct setent_ctx);
+ if (!req) {
+ DEBUG(0, ("Could not create tevent request for setgrent\n"));
+ return NULL;
+ }
+
+ state->nctx = nctx;
+ state->client = client;
+
+ /* do not query backends if we have a recent enumeration */
+ if (nctx->enum_cache_timeout) {
+ if (nctx->last_user_enum + nctx->enum_cache_timeout > now) {
+ enum_cached = true;
}
- nctx->gctx = gctx;
}
+ state->dctx = talloc_zero(state, struct nss_dom_ctx);
+ if (!state->dctx) {
+ ret = ENOMEM;
+ goto error;
+ }
+
+ /* check if enumeration is enabled in any domain */
+ for (dom = client->rctx->domains; dom; dom = dom->next) {
+ if (dom->enumerate != 0) break;
+ }
+ state->dctx->domain = dom;
+
+ if (state->dctx->domain == NULL) {
+ DEBUG(2, ("Enumeration disabled on all domains!\n"));
+ ret = ENOENT;
+ goto error;
+ }
+
+ state->dctx->check_provider =
+ NEED_CHECK_PROVIDER(state->dctx->domain->provider);
+
+ /* Is the result context already available */
+ if (state->nctx->gctx) {
+ if (state->nctx->gctx->ready) {
+ /* All of the necessary data is in place
+ * We can return now, getgrent requests will work at this point
+ */
+ tevent_req_done(req);
+ tevent_req_post(req, state->nctx->rctx->ev);
+ }
+ else {
+ /* Object is still being constructed
+ * Register for notification when it's
+ * ready.
+ */
+ ret = setent_add_ref(state->client, state->nctx->gctx, req);
+ if (ret != EOK) {
+ talloc_free(req);
+ return NULL;
+ }
+ }
+ return req;
+ }
+
+ /* Create a new result context
+ * We are creating it on the nss_ctx so that it doesn't
+ * go away if the original request does. We will delete
+ * it when the refcount goes to zero;
+ */
+ state->nctx->gctx = talloc_zero(nctx, struct getent_ctx);
+ if (!state->nctx->gctx) {
+ ret = ENOMEM;
+ goto error;
+ }
+ state->getent_ctx = nctx->gctx;
+
+ /* Add a callback reference for ourselves */
+ setent_add_ref(state->client, state->nctx->gctx, req);
+
+ /* ok, start the searches */
+ step_ctx = talloc_zero(state->getent_ctx, struct setent_step_ctx);
+ if (!step_ctx) {
+ ret = ENOMEM;
+ goto error;
+ }
+
+ /* Steal the dom_ctx onto the step_ctx so it doesn't go out of scope if
+ * this request is canceled while other requests are in-progress.
+ */
+ step_ctx->dctx = talloc_steal(step_ctx, state->dctx);
+ step_ctx->nctx = state->nctx;
+ step_ctx->getent_ctx = state->getent_ctx;
+ step_ctx->rctx = client->rctx;
+ step_ctx->enum_cached = enum_cached;
+
+ ret = nss_cmd_setgrent_step(step_ctx);
+ if (ret != EOK) goto error;
+
+ return req;
+
+ error:
+ tevent_req_error(req, ret);
+ tevent_req_post(req, client->rctx->ev);
+ return req;
+}
+
+static void nss_cmd_setgrent_dp_callback(uint16_t err_maj, uint32_t err_min,
+ const char *err_msg, void *ptr);
+static void setgrent_result_timeout(struct tevent_context *ev,
+ struct tevent_timer *te,
+ struct timeval current_time,
+ void *pvt);
+static errno_t nss_cmd_setgrent_step(struct setent_step_ctx *step_ctx)
+{
+ errno_t ret;
+ struct sss_domain_info *dom = step_ctx->dctx->domain;
+ struct resp_ctx *rctx = step_ctx->rctx;
+ struct nss_dom_ctx *dctx = step_ctx->dctx;
+ struct getent_ctx *gctx = step_ctx->getent_ctx;
+ struct nss_ctx *nctx = step_ctx->nctx;
+ struct sysdb_ctx *sysdb;
+ struct ldb_result *res;
+ struct setent_req_list *req;
+ struct timeval tv;
+ struct tevent_timer *te;
+ int timeout;
+
while (dom) {
while (dom && dom->enumerate == 0) {
dom = dom->next;
@@ -2051,7 +2433,7 @@ static int nss_cmd_getgrent_search(struct nss_dom_ctx *dctx)
if (dom != dctx->domain) {
/* make sure we reset the check_provider flag when we check
* a new domain */
- if (cmdctx->enum_cached) {
+ if (step_ctx->enum_cached) {
dctx->check_provider = false;
} else {
dctx->check_provider = NEED_CHECK_PROVIDER(dom->provider);
@@ -2061,9 +2443,9 @@ static int nss_cmd_getgrent_search(struct nss_dom_ctx *dctx)
/* make sure to update the dctx if we changed domain */
dctx->domain = dom;
- DEBUG(4, ("Requesting info for domain [%s]\n", dom->name));
+ DEBUG(6, ("Requesting info for domain [%s]\n", dom->name));
- ret = sysdb_get_ctx_from_list(cctx->rctx->db_list, dom, &sysdb);
+ ret = sysdb_get_ctx_from_list(rctx->db_list, dom, &sysdb);
if (ret != EOK) {
DEBUG(0, ("Fatal: Sysdb CTX not found for this domain!\n"));
return EIO;
@@ -2072,10 +2454,12 @@ static int nss_cmd_getgrent_search(struct nss_dom_ctx *dctx)
/* if this is a caching provider (or if we haven't checked the cache
* yet) then verify that the cache is uptodate */
if (dctx->check_provider) {
+ /* Only do this once per provider */
dctx->check_provider = false;
timeout = SSS_CLI_SOCKET_TIMEOUT;
- ret = sss_dp_send_acct_req(cctx->rctx, cmdctx,
- nss_cmd_getgrent_dp_callback, dctx,
+
+ ret = sss_dp_send_acct_req(rctx, step_ctx,
+ nss_cmd_setgrent_dp_callback, step_ctx,
timeout, dom->name, true,
SSS_DP_GROUP, NULL, 0);
if (ret == EOK) {
@@ -2101,8 +2485,7 @@ static int nss_cmd_getgrent_search(struct nss_dom_ctx *dctx)
continue;
}
-
- gctx->doms = talloc_realloc(gctx, gctx->doms,
+ nctx->gctx->doms = talloc_realloc(gctx, gctx->doms,
struct dom_ctx, gctx->num +1);
if (!gctx->doms) {
talloc_free(gctx);
@@ -2110,41 +2493,65 @@ static int nss_cmd_getgrent_search(struct nss_dom_ctx *dctx)
return ENOMEM;
}
- gctx->doms[gctx->num].domain = dctx->domain;
- gctx->doms[gctx->num].res = talloc_steal(gctx->doms, res);
- gctx->doms[gctx->num].cur = 0;
+ nctx->gctx->doms[gctx->num].domain = dctx->domain;
+ nctx->gctx->doms[gctx->num].res = talloc_steal(gctx->doms, res);
- gctx->num++;
+ nctx->gctx->num++;
/* do not reply until all domain searches are done */
dom = dom->next;
}
- /* set cache mark */
- nctx->last_group_enum = time(NULL);
+ /* We've finished all our lookups
+ * The result object is now safe to read.
+ */
+ nctx->gctx->ready = true;
- if (cmdctx->immediate) {
- /* this was a getgrent call w/o setgrent,
- * return immediately one result */
- return nss_cmd_getgrent_immediate(cmdctx);
+ /* Set up a lifetime timer for this result object
+ * We don't want this result object to outlive the
+ * enum cache refresh timeout
+ */
+ tv = tevent_timeval_current_ofs(nctx->enum_cache_timeout, 0);
+ te = tevent_add_timer(rctx->ev, nctx->gctx, tv,
+ setgrent_result_timeout, nctx);
+ if (!te) {
+ DEBUG(0, ("Could not set up life timer for setgrent result object. "
+ "Entries may become stale.\n"));
}
- /* create response packet */
- ret = sss_packet_new(cctx->creq, 0,
- sss_packet_get_cmd(cctx->creq->in),
- &cctx->creq->out);
- if (ret == EOK) {
- sss_cmd_done(cctx, cmdctx);
+ /* Notify the waiting clients */
+ while (nctx->gctx->reqs) {
+ tevent_req_done(nctx->gctx->reqs->req);
+ /* Freeing each entry in the list removes it from the dlist */
+ req = nctx->gctx->reqs;
+ nctx->gctx->reqs = nctx->gctx->reqs->next;
+ talloc_free(req);
}
- return ret;
+ return EOK;
}
-static void nss_cmd_getgrent_dp_callback(uint16_t err_maj, uint32_t err_min,
+static void setgrent_result_timeout(struct tevent_context *ev,
+ struct tevent_timer *te,
+ struct timeval current_time,
+ void *pvt)
+{
+ struct nss_ctx *nctx = talloc_get_type(pvt, struct nss_ctx);
+
+ DEBUG(1, ("setgrent result object has expired. Cleaning up.\n"));
+
+ /* Free the group enumeration context.
+ * If additional getgrent requests come in, they will invoke
+ * an implicit setgrent and refresh the result object.
+ */
+ talloc_zfree(nctx->gctx);
+}
+
+static void nss_cmd_setgrent_dp_callback(uint16_t err_maj, uint32_t err_min,
const char *err_msg, void *ptr)
{
- struct nss_dom_ctx *dctx = talloc_get_type(ptr, struct nss_dom_ctx);
- struct nss_cmd_ctx *cmdctx = dctx->cmdctx;
- struct cli_ctx *cctx = cmdctx->cctx;
+ struct setent_step_ctx *step_ctx =
+ talloc_get_type(ptr, struct setent_step_ctx);
+ struct getent_ctx *gctx = step_ctx->nctx->gctx;
int ret;
if (err_maj) {
@@ -2154,81 +2561,44 @@ static void nss_cmd_getgrent_dp_callback(uint16_t err_maj, uint32_t err_min,
(unsigned int)err_maj, (unsigned int)err_min, err_msg));
}
- ret = nss_cmd_getgrent_search(dctx);
-
+ ret = nss_cmd_setgrent_step(step_ctx);
if (ret) {
- NSS_CMD_FATAL_ERROR(cctx);
+ /* Notify any waiting processes of failure */
+ while(step_ctx->nctx->gctx->reqs) {
+ tevent_req_error(gctx->reqs->req, ret);
+ /* Freeing each entry in the list removes it from the dlist */
+ talloc_free(gctx->reqs);
+ }
}
}
-static int nss_cmd_setgrent_ext(struct cli_ctx *cctx, bool immediate)
+static errno_t nss_cmd_setgrent_recv(struct tevent_req *req)
{
- struct sss_domain_info *dom;
- struct nss_cmd_ctx *cmdctx;
- struct nss_dom_ctx *dctx;
- struct nss_ctx *nctx;
- time_t now = time(NULL);
- int ret;
-
- DEBUG(4, ("Requesting info for all groups\n"));
-
- nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
- talloc_free(nctx->gctx);
- nctx->gctx = NULL;
-
- cmdctx = talloc_zero(cctx, struct nss_cmd_ctx);
- if (!cmdctx) {
- return ENOMEM;
- }
- cmdctx->cctx = cctx;
- cmdctx->immediate = immediate;
-
- dctx = talloc_zero(cmdctx, struct nss_dom_ctx);
- if (!dctx) {
- ret = ENOMEM;
- goto done;
- }
- dctx->cmdctx = cmdctx;
-
- /* do not query backends if we have a recent enumeration */
- if (nctx->enum_cache_timeout) {
- if (nctx->last_group_enum + nctx->enum_cache_timeout > now) {
- cmdctx->enum_cached = true;
- }
- }
+ TEVENT_REQ_RETURN_ON_ERROR(req);
+ return EOK;
+}
- /* check if enumeration is enabled in any domain */
- for (dom = cctx->rctx->domains; dom; dom = dom->next) {
- if (dom->enumerate != 0) break;
- }
- dctx->domain = dom;
+static void nss_cmd_setgrent_done(struct tevent_req *req)
+{
+ errno_t ret;
+ struct nss_cmd_ctx *cmdctx =
+ tevent_req_callback_data(req, struct nss_cmd_ctx);
- if (dctx->domain == NULL) {
- DEBUG(2, ("Enumeration disabled on all domains!\n"));
- if (cmdctx->immediate) {
- ret = ENOENT;
- } else {
- ret = sss_packet_new(cctx->creq, 0,
- sss_packet_get_cmd(cctx->creq->in),
- &cctx->creq->out);
- if (ret == EOK) {
- sss_cmd_done(cctx, cmdctx);
- }
+ ret = nss_cmd_setgrent_recv(req);
+ talloc_zfree(req);
+ if (ret == EOK || ret == ENOENT) {
+ /* Either we succeeded or no domains were eligible */
+ ret = sss_packet_new(cmdctx->cctx->creq, 0,
+ sss_packet_get_cmd(cmdctx->cctx->creq->in),
+ &cmdctx->cctx->creq->out);
+ if (ret == EOK) {
+ sss_cmd_done(cmdctx->cctx, cmdctx);
+ return;
}
- goto done;
}
- dctx->check_provider = NEED_CHECK_PROVIDER(dctx->domain->provider);
-
- /* ok, start the searches */
- ret = nss_cmd_getgrent_search(dctx);
-done:
- return nss_cmd_done(cmdctx, ret);
-}
-
-static int nss_cmd_setgrent(struct cli_ctx *cctx)
-{
- return nss_cmd_setgrent_ext(cctx, false);
+ /* Something bad happened */
+ nss_cmd_done(cmdctx, ret);
}
static int nss_cmd_retgrent(struct cli_ctx *cctx, int num)
@@ -2246,28 +2616,28 @@ static int nss_cmd_retgrent(struct cli_ctx *cctx, int num)
gctx = nctx->gctx;
while (ret == ENOENT) {
- if (gctx->cur >= gctx->num) break;
+ if (cctx->grent_dom_idx >= gctx->num) break;
- gdom = &gctx->doms[gctx->cur];
+ gdom = &gctx->doms[cctx->grent_dom_idx];
- n = gdom->res->count - gdom->cur;
- if (n == 0 && (gctx->cur+1 < gctx->num)) {
- gctx->cur++;
- gdom = &gctx->doms[gctx->cur];
- n = gdom->res->count - gdom->cur;
+ n = gdom->res->count - cctx->grent_cur;
+ if (n <= 0 && (cctx->grent_cur+1 < gctx->num)) {
+ cctx->grent_dom_idx++;
+ gdom = &gctx->doms[cctx->grent_dom_idx];
+ n = gdom->res->count - cctx->grent_cur;
}
if (!n) break;
if (n > num) n = num;
- msgs = &(gdom->res->msgs[gdom->cur]);
+ msgs = &(gdom->res->msgs[cctx->grent_cur]);
ret = fill_grent(cctx->creq->out,
gdom->domain,
nctx, true, msgs, &n);
- gdom->cur += n;
+ cctx->grent_cur += n;
}
none:
@@ -2308,29 +2678,77 @@ static int nss_cmd_getgrent_immediate(struct nss_cmd_ctx *cmdctx)
return EOK;
}
+static void nss_cmd_implicit_setgrent_done(struct tevent_req *req);
static int nss_cmd_getgrent(struct cli_ctx *cctx)
{
struct nss_ctx *nctx;
struct nss_cmd_ctx *cmdctx;
+ struct tevent_req *req;
DEBUG(4, ("Requesting info for all groups\n"));
- nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
-
- /* see if we need to trigger an implicit setpwent() */
- if (nctx->gctx == NULL) {
- return nss_cmd_setgrent_ext(cctx, true);
- }
-
cmdctx = talloc(cctx, struct nss_cmd_ctx);
if (!cmdctx) {
return ENOMEM;
}
cmdctx->cctx = cctx;
+ /* Save the current index and cursor locations
+ * If we end up calling setgrent implicitly, because the response object
+ * expired and has to be recreated, we want to resume from the same
+ * location.
+ */
+ cmdctx->saved_dom_idx = cctx->grent_dom_idx;
+ cmdctx->saved_cur = cctx->grent_cur;
+
+ nctx = talloc_get_type(cctx->rctx->pvt_ctx, struct nss_ctx);
+ if(!nctx->gctx || !nctx->gctx->ready) {
+ /* Make sure we invoke setgrent if it hasn't been run or is still
+ * processing from another client
+ */
+ req = nss_cmd_setgrent_send(cctx, cctx);
+ if (!req) {
+ return EIO;
+ }
+ tevent_req_set_callback(req, nss_cmd_implicit_setgrent_done, cmdctx);
+ return EOK;
+ }
+
return nss_cmd_getgrent_immediate(cmdctx);
}
+static errno_t nss_cmd_setgrent_recv(struct tevent_req *req);
+static void nss_cmd_implicit_setgrent_done(struct tevent_req *req)
+{
+ errno_t ret;
+ struct nss_cmd_ctx *cmdctx =
+ tevent_req_callback_data(req, struct nss_cmd_ctx);
+
+ ret = nss_cmd_setgrent_recv(req);
+ talloc_zfree(req);
+
+ /* ENOENT is acceptable, as it just means that there were no entries
+ * to be returned. This will be handled gracefully in nss_cmd_retpwent
+ * later.
+ */
+ if (ret != EOK && ret != ENOENT) {
+ DEBUG(0, ("Implicit setgrent failed with unexpected error [%d][%s]\n",
+ ret, strerror(ret)));
+ NSS_CMD_FATAL_ERROR(cmdctx);
+ }
+
+ /* Restore the saved index and cursor locations */
+ cmdctx->cctx->grent_dom_idx = cmdctx->saved_dom_idx;
+ cmdctx->cctx->grent_cur = cmdctx->saved_cur;
+
+ ret = nss_cmd_getgrent_immediate(cmdctx);
+ if (ret != EOK) {
+ DEBUG(0, ("Immediate retrieval failed with unexpected error "
+ "[%d][%s]\n", ret, strerror(ret)));
+ NSS_CMD_FATAL_ERROR(cmdctx);
+ }
+}
+
static int nss_cmd_endgrent(struct cli_ctx *cctx)
{
struct nss_ctx *nctx;
@@ -2350,9 +2768,9 @@ static int nss_cmd_endgrent(struct cli_ctx *cctx)
}
if (nctx->gctx == NULL) goto done;
- /* free results and reset */
- talloc_free(nctx->gctx);
- nctx->gctx = NULL;
+ /* Reset the indices so that subsequent requests start at zero */
+ cctx->grent_dom_idx = 0;
+ cctx->grent_cur = 0;
done:
sss_cmd_done(cctx, NULL);
@@ -2674,4 +3092,3 @@ int nss_cmd_execute(struct cli_ctx *cctx)
return EINVAL;
}
-