From: Jason Gunthorpe <jgg@nvidia.com> mainline inclusion from mainline-v5.15-rc4 commit 305d568b72f17f674155a2a8275f865f207b3808 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I9R4OB CVE: CVE-2021-47391 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... --------------------------- The FSM can run in a circle allowing rdma_resolve_ip() to be called twice on the same id_priv. While this cannot happen without going through the work, it violates the invariant that the same address resolution background request cannot be active twice. CPU 1 CPU 2 rdma_resolve_addr(): RDMA_CM_IDLE -> RDMA_CM_ADDR_QUERY rdma_resolve_ip(addr_handler) #1 process_one_req(): for #1 addr_handler(): RDMA_CM_ADDR_QUERY -> RDMA_CM_ADDR_BOUND mutex_unlock(&id_priv->handler_mutex); [.. handler still running ..] rdma_resolve_addr(): RDMA_CM_ADDR_BOUND -> RDMA_CM_ADDR_QUERY rdma_resolve_ip(addr_handler) !! two requests are now on the req_list rdma_destroy_id(): destroy_id_handler_unlock(): _destroy_id(): cma_cancel_operation(): rdma_addr_cancel() // process_one_req() self removes it spin_lock_bh(&lock); cancel_delayed_work(&req->work); if (!list_empty(&req->list)) == true ! rdma_addr_cancel() returns after process_on_req #1 is done kfree(id_priv) process_one_req(): for #2 addr_handler(): mutex_lock(&id_priv->handler_mutex); !! Use after free on id_priv rdma_addr_cancel() expects there to be one req on the list and only cancels the first one. The self-removal behavior of the work only happens after the handler has returned. This yields a situations where the req_list can have two reqs for the same "handle" but rdma_addr_cancel() only cancels the first one. The second req remains active beyond rdma_destroy_id() and will use-after-free id_priv once it inevitably triggers. Fix this by remembering if the id_priv has called rdma_resolve_ip() and always cancel before calling it again. This ensures the req_list never gets more than one item in it and doesn't cost anything in the normal flow that never uses this strange error path. Link: https://lore.kernel.org/r/0-v1-3bc675b8006d+22-syz_cancel_uaf_jgg@nvidia.com Cc: stable@vger.kernel.org Fixes: e51060f08a61 ("IB: IP address based RDMA connection manager") Reported-by: syzbot+dc3dfba010d7671e05f5@syzkaller.appspotmail.com Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> (cherry picked from commit 305d568b72f17f674155a2a8275f865f207b3808) Signed-off-by: Liu Jian <liujian56@huawei.com> Conflicts: drivers/infiniband/core/cma.c drivers/infiniband/core/cma_priv.h [Some contexts different. No functional impact.] Signed-off-by: Zheng Zucheng <zhengzucheng@huawei.com> --- drivers/infiniband/core/cma.c | 23 +++++++++++++++++++++++ drivers/infiniband/core/cma_priv.h | 1 + 2 files changed, 24 insertions(+) diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 20a5501b740a..90cb1255ca64 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -1643,6 +1643,14 @@ static void cma_cancel_operation(struct rdma_id_private *id_priv, { switch (state) { case RDMA_CM_ADDR_QUERY: + /* + * We can avoid doing the rdma_addr_cancel() based on state, + * only RDMA_CM_ADDR_QUERY has a work that could still execute. + * Notice that the addr_handler work could still be exiting + * outside this state, however due to the interaction with the + * handler_mutex the work is guaranteed not to touch id_priv + * during exit. + */ rdma_addr_cancel(&id_priv->id.route.addr.dev_addr); break; case RDMA_CM_ROUTE_QUERY: @@ -3459,6 +3467,21 @@ int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, if (dst_addr->sa_family == AF_IB) { ret = cma_resolve_ib_addr(id_priv); } else { + /* + * The FSM can return back to RDMA_CM_ADDR_BOUND after + * rdma_resolve_ip() is called, eg through the error + * path in addr_handler(). If this happens the existing + * request must be canceled before issuing a new one. + * Since canceling a request is a bit slow and this + * oddball path is rare, keep track once a request has + * been issued. The track turns out to be a permanent + * state since this is the only cancel as it is + * immediately before rdma_resolve_ip(). + */ + if (id_priv->used_resolve_ip) + rdma_addr_cancel(&id->route.addr.dev_addr); + else + id_priv->used_resolve_ip = 1; ret = rdma_resolve_ip(cma_src_addr(id_priv), dst_addr, &id->route.addr.dev_addr, timeout_ms, addr_handler, id_priv); diff --git a/drivers/infiniband/core/cma_priv.h b/drivers/infiniband/core/cma_priv.h index 194cfe78c447..e2e0bc89d847 100644 --- a/drivers/infiniband/core/cma_priv.h +++ b/drivers/infiniband/core/cma_priv.h @@ -87,6 +87,7 @@ struct rdma_id_private { bool tos_set; u8 reuseaddr; u8 afonly; + u8 used_resolve_ip; enum ib_gid_type gid_type; /* -- 2.34.1