Loading ifs/scripts/compile_dep.sh +5 −0 Original line number Diff line number Diff line Loading @@ -38,6 +38,8 @@ prepare_build_dir() { fi rm -rf $1/build/* } PATCH_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )" PATCH_DIR="${PATCH_DIR}/patches" CLUSTER="" NA_LAYER="" CORES="" Loading Loading @@ -204,6 +206,9 @@ if [ "$NA_LAYER" == "cci" ] || [ "$NA_LAYER" == "all" ]; then CURR=${SOURCE}/cci prepare_build_dir ${CURR} cd ${CURR} # patch hanging issue echo "########## Applying cci hanging patch" git apply ${PATCH_DIR}/cci_hang_final.patch || exit 1 ./autogen.pl || exit 1 cd ${CURR}/build if [ "$CLUSTER" == "mogon1" ]; then Loading ifs/scripts/dl_dep.sh +8 −3 Original line number Diff line number Diff line Loading @@ -174,14 +174,19 @@ if [ "${NA_LAYER}" == "ofi" ] || [ "${NA_LAYER}" == "all" ]; then clonedeps "libfabric" "git clone https://github.com/ofiwg/libfabric" "tags/v1.5.2" fi # get Mercury clonedeps "mercury" "git clone --recurse-submodules https://github.com/mercury-hpc/mercury" "afd70055d21a6df2faefe38d5f6ce1ae11f365a5" clonedeps "mercury" "git clone --recurse-submodules https://github.com/mercury-hpc/mercury" "2a7369db11c7d9e962a8d59f1852a04c4ff57694" # get Argobots clonedeps "argobots" "git clone -b dev-get-dev-basic https://github.com/carns/argobots.git" "a5a6b2036c75ad05804ccb72d2fe31cea1bfef88" clonedeps "argobots" "git clone -b dev-get-dev-basic https://github.com/carns/argobots.git" "fd6ae0f4613187a2c73fceee8a2718d54719bcab" # get Argobots-snoozer clonedeps "abt-snoozer" "git clone https://xgitlab.cels.anl.gov/sds/abt-snoozer.git" "3d9240eda290bfb89f08a5673cebd888194a4bd7" # get Margo clonedeps "margo" "git clone https://xgitlab.cels.anl.gov/sds/margo.git" "68ef7f14178e9066cf38846d90d451e00aaca61d" clonedeps "margo" "git clone https://xgitlab.cels.anl.gov/sds/margo.git" "30521c87a5d6b531b894877f40f0f34ef8ecd541" # get rocksdb clonedeps "rocksdb" "git clone https://github.com/facebook/rocksdb" "tags/v5.8" # Commit values used before 17.01.2018 # Mercury: afd70055d21a6df2faefe38d5f6ce1ae11f365a5 # Argobots: a5a6b2036c75ad05804ccb72d2fe31cea1bfef88 # Margo: 68ef7f14178e9066cf38846d90d451e00aaca61d echo "Nothing left to do. Exiting." ifs/scripts/patches/cci_hang_final.patch 0 → 100644 +21 −0 Original line number Diff line number Diff line diff --git a/src/plugins/ctp/verbs/ctp_verbs_api.c b/src/plugins/ctp/verbs/ctp_verbs_api.c index e187ebd..016d97b 100644 --- a/src/plugins/ctp/verbs/ctp_verbs_api.c +++ b/src/plugins/ctp/verbs/ctp_verbs_api.c @@ -1063,7 +1063,7 @@ void rdma_destroy_ep(struct rdma_cm_id *id) if (id->qp) rdma_destroy_qp(id); - rdma_destroy_id(id); + //rdma_destroy_id(id); } #endif /* HAVE_RDMA_ADDRINFO */ @@ -2363,6 +2363,7 @@ static void verbs_destroy_conn(cci__ep_t *ep, cci__conn_t *conn) __func__, strerror(ret)); } } + rdma_destroy_ep(vconn->id); pthread_rwlock_wrlock(&vep->conn_tree_lock); tdelete(&vconn->qp_num, &vep->conn_tree, verbs_compare_u32); Loading
ifs/scripts/compile_dep.sh +5 −0 Original line number Diff line number Diff line Loading @@ -38,6 +38,8 @@ prepare_build_dir() { fi rm -rf $1/build/* } PATCH_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )" PATCH_DIR="${PATCH_DIR}/patches" CLUSTER="" NA_LAYER="" CORES="" Loading Loading @@ -204,6 +206,9 @@ if [ "$NA_LAYER" == "cci" ] || [ "$NA_LAYER" == "all" ]; then CURR=${SOURCE}/cci prepare_build_dir ${CURR} cd ${CURR} # patch hanging issue echo "########## Applying cci hanging patch" git apply ${PATCH_DIR}/cci_hang_final.patch || exit 1 ./autogen.pl || exit 1 cd ${CURR}/build if [ "$CLUSTER" == "mogon1" ]; then Loading
ifs/scripts/dl_dep.sh +8 −3 Original line number Diff line number Diff line Loading @@ -174,14 +174,19 @@ if [ "${NA_LAYER}" == "ofi" ] || [ "${NA_LAYER}" == "all" ]; then clonedeps "libfabric" "git clone https://github.com/ofiwg/libfabric" "tags/v1.5.2" fi # get Mercury clonedeps "mercury" "git clone --recurse-submodules https://github.com/mercury-hpc/mercury" "afd70055d21a6df2faefe38d5f6ce1ae11f365a5" clonedeps "mercury" "git clone --recurse-submodules https://github.com/mercury-hpc/mercury" "2a7369db11c7d9e962a8d59f1852a04c4ff57694" # get Argobots clonedeps "argobots" "git clone -b dev-get-dev-basic https://github.com/carns/argobots.git" "a5a6b2036c75ad05804ccb72d2fe31cea1bfef88" clonedeps "argobots" "git clone -b dev-get-dev-basic https://github.com/carns/argobots.git" "fd6ae0f4613187a2c73fceee8a2718d54719bcab" # get Argobots-snoozer clonedeps "abt-snoozer" "git clone https://xgitlab.cels.anl.gov/sds/abt-snoozer.git" "3d9240eda290bfb89f08a5673cebd888194a4bd7" # get Margo clonedeps "margo" "git clone https://xgitlab.cels.anl.gov/sds/margo.git" "68ef7f14178e9066cf38846d90d451e00aaca61d" clonedeps "margo" "git clone https://xgitlab.cels.anl.gov/sds/margo.git" "30521c87a5d6b531b894877f40f0f34ef8ecd541" # get rocksdb clonedeps "rocksdb" "git clone https://github.com/facebook/rocksdb" "tags/v5.8" # Commit values used before 17.01.2018 # Mercury: afd70055d21a6df2faefe38d5f6ce1ae11f365a5 # Argobots: a5a6b2036c75ad05804ccb72d2fe31cea1bfef88 # Margo: 68ef7f14178e9066cf38846d90d451e00aaca61d echo "Nothing left to do. Exiting."
ifs/scripts/patches/cci_hang_final.patch 0 → 100644 +21 −0 Original line number Diff line number Diff line diff --git a/src/plugins/ctp/verbs/ctp_verbs_api.c b/src/plugins/ctp/verbs/ctp_verbs_api.c index e187ebd..016d97b 100644 --- a/src/plugins/ctp/verbs/ctp_verbs_api.c +++ b/src/plugins/ctp/verbs/ctp_verbs_api.c @@ -1063,7 +1063,7 @@ void rdma_destroy_ep(struct rdma_cm_id *id) if (id->qp) rdma_destroy_qp(id); - rdma_destroy_id(id); + //rdma_destroy_id(id); } #endif /* HAVE_RDMA_ADDRINFO */ @@ -2363,6 +2363,7 @@ static void verbs_destroy_conn(cci__ep_t *ep, cci__conn_t *conn) __func__, strerror(ret)); } } + rdma_destroy_ep(vconn->id); pthread_rwlock_wrlock(&vep->conn_tree_lock); tdelete(&vconn->qp_num, &vep->conn_tree, verbs_compare_u32);