| Message ID | 20220218003543.205799-2-rpearsonhpe@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Move two object pools to rxe_mcast.c | expand |
On Thu, Feb 17, 2022 at 06:35:39PM -0600, Bob Pearson wrote: > Well behaved applications will free all memory allocated by multicast. > If this fails rdma-core will detach all qp's from multicast groups when > an application terminates. This patch prints a warning and cleans out > memory allocated by multicast if both of these fail to occur. > > Signed-off-by: Bob Pearson <rpearsonhpe@gmail.com> > drivers/infiniband/sw/rxe/rxe.c | 2 ++ > drivers/infiniband/sw/rxe/rxe_loc.h | 1 + > drivers/infiniband/sw/rxe/rxe_mcast.c | 38 +++++++++++++++++++++++++++ > 3 files changed, 41 insertions(+) > > diff --git a/drivers/infiniband/sw/rxe/rxe.c b/drivers/infiniband/sw/rxe/rxe.c > index 3520eb2db685..603b0156f889 100644 > +++ b/drivers/infiniband/sw/rxe/rxe.c > @@ -29,6 +29,8 @@ void rxe_dealloc(struct ib_device *ib_dev) > rxe_pool_cleanup(&rxe->mr_pool); > rxe_pool_cleanup(&rxe->mw_pool); > > + rxe_cleanup_mcast(rxe); Lets just put WARN_ON(!RB_EMPTY_ROOT(&rxe->mcg_tree)); Righ here and forget about there rest. Leaking memory if the kernel is buggy is OK. Jason
diff --git a/drivers/infiniband/sw/rxe/rxe.c b/drivers/infiniband/sw/rxe/rxe.c index 3520eb2db685..603b0156f889 100644 --- a/drivers/infiniband/sw/rxe/rxe.c +++ b/drivers/infiniband/sw/rxe/rxe.c @@ -29,6 +29,8 @@ void rxe_dealloc(struct ib_device *ib_dev) rxe_pool_cleanup(&rxe->mr_pool); rxe_pool_cleanup(&rxe->mw_pool); + rxe_cleanup_mcast(rxe); + if (rxe->tfm) crypto_free_shash(rxe->tfm); } diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h b/drivers/infiniband/sw/rxe/rxe_loc.h index 409efeecd581..0bc1b7e2877c 100644 --- a/drivers/infiniband/sw/rxe/rxe_loc.h +++ b/drivers/infiniband/sw/rxe/rxe_loc.h @@ -44,6 +44,7 @@ struct rxe_mcg *rxe_lookup_mcg(struct rxe_dev *rxe, union ib_gid *mgid); int rxe_attach_mcast(struct ib_qp *ibqp, union ib_gid *mgid, u16 mlid); int rxe_detach_mcast(struct ib_qp *ibqp, union ib_gid *mgid, u16 mlid); void rxe_cleanup_mcg(struct kref *kref); +void rxe_cleanup_mcast(struct rxe_dev *rxe); /* rxe_mmap.c */ struct rxe_mmap_info { diff --git a/drivers/infiniband/sw/rxe/rxe_mcast.c b/drivers/infiniband/sw/rxe/rxe_mcast.c index 4935fe5c5868..447d78bea28b 100644 --- a/drivers/infiniband/sw/rxe/rxe_mcast.c +++ b/drivers/infiniband/sw/rxe/rxe_mcast.c @@ -388,3 +388,41 @@ int rxe_detach_mcast(struct ib_qp *ibqp, union ib_gid *mgid, u16 mlid) return rxe_detach_mcg(rxe, qp, mgid); } + +/** + * rxe_cleanup_mcast - cleanup all resources still held by mcast + * @rxe: rxe object + * + * Called when rxe device is unloaded. Walk red-black tree to + * find all mcg's and then walk mcg->qp_list to find all mca's and + * free them. + * + * This is belt and suspenders. These should have been recovered + * already by the application else rdma-core. + */ +void rxe_cleanup_mcast(struct rxe_dev *rxe) +{ + struct rb_root *root = &rxe->mcg_tree; + struct rb_node *node, *next; + struct rxe_mcg *mcg; + struct rxe_mca *mca, *tmp; + + if (RB_EMPTY_ROOT(root)) + return; /* expected behavior */ + + pr_warn("RDMA/core bug: Recovering leaked multicast resources\n"); + + for (node = rb_first(root); node; node = next) { + next = rb_next(node); + mcg = rb_entry(node, typeof(*mcg), node); + + spin_lock_bh(&rxe->mcg_lock); + list_for_each_entry_safe(mca, tmp, &mcg->qp_list, qp_list) + kfree(mca); + + __rxe_remove_mcg(mcg); + spin_unlock_bh(&rxe->mcg_lock); + + kfree(mcg); + } +}
Well behaved applications will free all memory allocated by multicast. If this fails rdma-core will detach all qp's from multicast groups when an application terminates. This patch prints a warning and cleans out memory allocated by multicast if both of these fail to occur. Signed-off-by: Bob Pearson <rpearsonhpe@gmail.com> --- drivers/infiniband/sw/rxe/rxe.c | 2 ++ drivers/infiniband/sw/rxe/rxe_loc.h | 1 + drivers/infiniband/sw/rxe/rxe_mcast.c | 38 +++++++++++++++++++++++++++ 3 files changed, 41 insertions(+)