From 21a446cf186570168b7281b154b1993968598aca Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 5 Nov 2018 11:10:50 -0500 Subject: NFSv4: Don't exit the state manager without clearing NFS4CLNT_MANAGER_RUNNING If we exit the NFSv4 state manager due to a umount, then we can end up leaving the NFS4CLNT_MANAGER_RUNNING flag set. If another mount causes the nfs4_client to be rereferenced before it is destroyed, then we end up never being able to recover state. Fixes: 47c2199b6eb5 ("NFSv4.1: Ensure state manager thread dies on last ...") Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org # v4.15+ --- fs/nfs/nfs4state.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'fs/nfs/nfs4state.c') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 62ae0fd345ad..98d1b6a6646a 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -2601,11 +2601,12 @@ static void nfs4_state_manager(struct nfs_client *clp) nfs4_clear_state_manager_bit(clp); /* Did we race with an attempt to give us more work? */ if (clp->cl_state == 0) - break; + return; if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0) - break; + return; } while (refcount_read(&clp->cl_count) > 1); - return; + goto out_drain; + out_error: if (strlen(section)) section_sep = ": "; @@ -2613,6 +2614,7 @@ out_error: " with error %d\n", section_sep, section, clp->cl_hostname, -status); ssleep(1); +out_drain: nfs4_end_drain_session(clp); nfs4_clear_state_manager_bit(clp); } -- cgit From a1aa09be21fa344d1f5585aab8164bfae55f57e3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 5 Nov 2018 12:17:01 -0500 Subject: NFSv4: Ensure that the state manager exits the loop on SIGKILL Signed-off-by: Trond Myklebust --- fs/nfs/nfs4state.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/nfs4state.c') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 98d1b6a6646a..ffea57885394 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -2604,7 +2604,7 @@ static void nfs4_state_manager(struct nfs_client *clp) return; if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0) return; - } while (refcount_read(&clp->cl_count) > 1); + } while (refcount_read(&clp->cl_count) > 1 && !signalled()); goto out_drain; out_error: -- cgit From aeabb3c96186a0f944fc2b1f25c84d5eb3a93fa9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 19 Nov 2018 20:11:45 -0500 Subject: NFSv4: Fix a NFSv4 state manager deadlock Fix a deadlock whereby the NFSv4 state manager can get stuck in the delegation return code, waiting for a layout return to complete in another thread. If the server reboots before that other thread completes, then we need to be able to start a second state manager thread in order to perform recovery. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4_fs.h | 2 ++ fs/nfs/nfs4state.c | 16 +++++++++++----- 2 files changed, 13 insertions(+), 5 deletions(-) (limited to 'fs/nfs/nfs4state.c') diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 8d59c9655ec4..1b994b527518 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -41,6 +41,8 @@ enum nfs4_client_state { NFS4CLNT_MOVED, NFS4CLNT_LEASE_MOVED, NFS4CLNT_DELEGATION_EXPIRED, + NFS4CLNT_RUN_MANAGER, + NFS4CLNT_DELEGRETURN_RUNNING, }; #define NFS4_RENEW_TIMEOUT 0x01 diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index ffea57885394..d8decf2ec48f 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1210,6 +1210,7 @@ void nfs4_schedule_state_manager(struct nfs_client *clp) struct task_struct *task; char buf[INET6_ADDRSTRLEN + sizeof("-manager") + 1]; + set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state); if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0) return; __module_get(THIS_MODULE); @@ -2503,6 +2504,7 @@ static void nfs4_state_manager(struct nfs_client *clp) /* Ensure exclusive access to NFSv4 state */ do { + clear_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state); if (test_bit(NFS4CLNT_PURGE_STATE, &clp->cl_state)) { section = "purge state"; status = nfs4_purge_lease(clp); @@ -2593,14 +2595,18 @@ static void nfs4_state_manager(struct nfs_client *clp) } nfs4_end_drain_session(clp); - if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) { - nfs_client_return_marked_delegations(clp); - continue; + nfs4_clear_state_manager_bit(clp); + + if (!test_and_set_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state)) { + if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) { + nfs_client_return_marked_delegations(clp); + set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state); + } + clear_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state); } - nfs4_clear_state_manager_bit(clp); /* Did we race with an attempt to give us more work? */ - if (clp->cl_state == 0) + if (!test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state)) return; if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0) return; -- cgit