Commit caa06154 authored by Seth Howell's avatar Seth Howell Committed by Ben Walker
Browse files

rdma: fix the poll_group_create error paths. 



It was possible to leak pollers if we had multiple devices in the
transport. The new err_exit path fixes this.

Change-Id: Iafd5643c67fae741113f10afe761af1988cb6a9b
Signed-off-by: default avatarSeth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/c/439419


Tested-by: default avatarSPDK CI Jenkins <sys_sgci@intel.com>
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: default avatarShuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: default avatarDarek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: default avatarBen Walker <benjamin.walker@intel.com>
parent 3709db73

lib/nvmf/rdma.c

+16 −7
Original line number Diff line number Diff line
@@ -2373,7 +2373,7 @@ spdk_nvmf_rdma_poll_group_create(struct spdk_nvmf_transport *transport) 
{

	struct spdk_nvmf_rdma_transport		*rtransport;

	struct spdk_nvmf_rdma_poll_group	*rgroup;

	struct spdk_nvmf_rdma_poller		*poller;

	struct spdk_nvmf_rdma_poller		*poller, *tpoller;

	struct spdk_nvmf_rdma_device		*device;



	rtransport = SPDK_CONTAINEROF(transport, struct spdk_nvmf_rdma_transport, transport);
@@ -2390,9 +2390,7 @@ spdk_nvmf_rdma_poll_group_create(struct spdk_nvmf_transport *transport) 
		poller = calloc(1, sizeof(*poller));

		if (!poller) {

			SPDK_ERRLOG("Unable to allocate memory for new RDMA poller\n");

			free(rgroup);

			pthread_mutex_unlock(&rtransport->lock);

			return NULL;

			goto err_exit;

		}



		poller->device = device;
@@ -2404,9 +2402,7 @@ spdk_nvmf_rdma_poll_group_create(struct spdk_nvmf_transport *transport) 
		if (!poller->cq) {

			SPDK_ERRLOG("Unable to create completion queue\n");

			free(poller);

			free(rgroup);

			pthread_mutex_unlock(&rtransport->lock);

			return NULL;

			goto err_exit;

		}



		TAILQ_INSERT_TAIL(&rgroup->pollers, poller, link);
@@ -2414,6 +2410,19 @@ spdk_nvmf_rdma_poll_group_create(struct spdk_nvmf_transport *transport) 


	pthread_mutex_unlock(&rtransport->lock);

	return &rgroup->group;



err_exit:

	TAILQ_FOREACH_SAFE(poller, &rgroup->pollers, link, tpoller) {

		TAILQ_REMOVE(&rgroup->pollers, poller, link);

		if (poller->cq) {

			ibv_destroy_cq(poller->cq);

		}

		free(poller);

	}



	free(rgroup);

	pthread_mutex_unlock(&rtransport->lock);

	return NULL;

}



static void