From 2a534a7473bf4e7f1c12805113f80c795fc8e89a Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 23 Aug 2018 11:02:49 -0400 Subject: NFSv4: Fix a tracepoint Oops in initiate_file_draining() Now that the value of 'ino' can be NULL or an ERR_PTR(), we need to change the test in the tracepoint. Fixes: ce5624f7e6675 ("NFSv4: Return NFS4ERR_DELAY when a layout fails...") Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org # v4.17+ Signed-off-by: Anna Schumaker --- fs/nfs/nfs4trace.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4trace.h b/fs/nfs/nfs4trace.h index a275fba93170..708342f4692f 100644 --- a/fs/nfs/nfs4trace.h +++ b/fs/nfs/nfs4trace.h @@ -1194,7 +1194,7 @@ DECLARE_EVENT_CLASS(nfs4_inode_stateid_callback_event, TP_fast_assign( __entry->error = error; __entry->fhandle = nfs_fhandle_hash(fhandle); - if (inode != NULL) { + if (!IS_ERR_OR_NULL(inode)) { __entry->fileid = NFS_FILEID(inode); __entry->dev = inode->i_sb->s_dev; } else { -- cgit v1.2.1 From d03360aaf5ccac49581960bd736258c62972b88b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 5 Sep 2018 14:07:12 -0400 Subject: pNFS: Ensure we return the error if someone kills a waiting layoutget If someone interrupts a wait on one or more outstanding layoutgets in pnfs_update_layout() then return the ERESTARTSYS/EINTR error. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index e8f232de484f..7d9a51e6b847 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1740,16 +1740,16 @@ static bool pnfs_within_mdsthreshold(struct nfs_open_context *ctx, return ret; } -static bool pnfs_prepare_to_retry_layoutget(struct pnfs_layout_hdr *lo) +static int pnfs_prepare_to_retry_layoutget(struct pnfs_layout_hdr *lo) { /* * send layoutcommit as it can hold up layoutreturn due to lseg * reference */ pnfs_layoutcommit_inode(lo->plh_inode, false); - return !wait_on_bit_action(&lo->plh_flags, NFS_LAYOUT_RETURN, + return wait_on_bit_action(&lo->plh_flags, NFS_LAYOUT_RETURN, nfs_wait_bit_killable, - TASK_UNINTERRUPTIBLE); + TASK_KILLABLE); } static void nfs_layoutget_begin(struct pnfs_layout_hdr *lo) @@ -1830,7 +1830,9 @@ pnfs_update_layout(struct inode *ino, } lookup_again: - nfs4_client_recover_expired_lease(clp); + lseg = ERR_PTR(nfs4_client_recover_expired_lease(clp)); + if (IS_ERR(lseg)) + goto out; first = false; spin_lock(&ino->i_lock); lo = pnfs_find_alloc_layout(ino, ctx, gfp_flags); @@ -1863,9 +1865,9 @@ lookup_again: if (list_empty(&lo->plh_segs) && atomic_read(&lo->plh_outstanding) != 0) { spin_unlock(&ino->i_lock); - if (wait_var_event_killable(&lo->plh_outstanding, - atomic_read(&lo->plh_outstanding) == 0 - || !list_empty(&lo->plh_segs))) + lseg = ERR_PTR(wait_var_event_killable(&lo->plh_outstanding, + atomic_read(&lo->plh_outstanding))); + if (IS_ERR(lseg) || !list_empty(&lo->plh_segs)) goto out_put_layout_hdr; pnfs_put_layout_hdr(lo); goto lookup_again; @@ -1898,8 +1900,11 @@ lookup_again: if (test_and_set_bit(NFS_LAYOUT_FIRST_LAYOUTGET, &lo->plh_flags)) { spin_unlock(&ino->i_lock); - wait_on_bit(&lo->plh_flags, NFS_LAYOUT_FIRST_LAYOUTGET, - TASK_UNINTERRUPTIBLE); + lseg = ERR_PTR(wait_on_bit(&lo->plh_flags, + NFS_LAYOUT_FIRST_LAYOUTGET, + TASK_KILLABLE)); + if (IS_ERR(lseg)) + goto out_put_layout_hdr; pnfs_put_layout_hdr(lo); dprintk("%s retrying\n", __func__); goto lookup_again; @@ -1925,7 +1930,8 @@ lookup_again: if (test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) { spin_unlock(&ino->i_lock); dprintk("%s wait for layoutreturn\n", __func__); - if (pnfs_prepare_to_retry_layoutget(lo)) { + lseg = ERR_PTR(pnfs_prepare_to_retry_layoutget(lo)); + if (!IS_ERR(lseg)) { if (first) pnfs_clear_first_layoutget(lo); pnfs_put_layout_hdr(lo); -- cgit v1.2.1 From 2edaead69e7573f35e8d5dc20938e41eacc21b35 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 5 Sep 2018 14:07:13 -0400 Subject: NFSv4: Fix a tracepoint Oops in initiate_file_draining() Now that the value of 'ino' can be NULL or an ERR_PTR(), we need to change the test in the tracepoint. Fixes: ce5624f7e6675 ("NFSv4: Return NFS4ERR_DELAY when a layout fails...") Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org # v4.17+ Signed-off-by: Anna Schumaker --- fs/nfs/nfs4trace.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4trace.h b/fs/nfs/nfs4trace.h index 708342f4692f..b1483b303e0b 100644 --- a/fs/nfs/nfs4trace.h +++ b/fs/nfs/nfs4trace.h @@ -1137,7 +1137,7 @@ DECLARE_EVENT_CLASS(nfs4_inode_callback_event, TP_fast_assign( __entry->error = error; __entry->fhandle = nfs_fhandle_hash(fhandle); - if (inode != NULL) { + if (!IS_ERR_OR_NULL(inode)) { __entry->fileid = NFS_FILEID(inode); __entry->dev = inode->i_sb->s_dev; } else { -- cgit v1.2.1 From 994b15b983a72e1148a173b61e5b279219bb45ae Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 5 Sep 2018 14:07:14 -0400 Subject: NFSv4.1 fix infinite loop on I/O. The previous fix broke recovery of delegated stateids because it assumes that if we did not mark the delegation as suspect, then the delegation has effectively been revoked, and so it removes that delegation irrespectively of whether or not it is valid and still in use. While this is "mostly harmless" for ordinary I/O, we've seen pNFS fail with LAYOUTGET spinning in an infinite loop while complaining that we're using an invalid stateid (in this case the all-zero stateid). What we rather want to do here is ensure that the delegation is always correctly marked as needing testing when that is the case. So we want to close the loophole offered by nfs4_schedule_stateid_recovery(), which marks the state as needing to be reclaimed, but not the delegation that may be backing it. Fixes: 0e3d3e5df07dc ("NFSv4.1 fix infinite loop on IO BAD_STATEID error") Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org # v4.11+ Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 10 +++++++--- fs/nfs/nfs4state.c | 2 ++ 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index df60dce935f3..094c3c09ff00 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2676,14 +2676,18 @@ static void nfs41_check_delegation_stateid(struct nfs4_state *state) } nfs4_stateid_copy(&stateid, &delegation->stateid); - if (test_bit(NFS_DELEGATION_REVOKED, &delegation->flags) || - !test_and_clear_bit(NFS_DELEGATION_TEST_EXPIRED, - &delegation->flags)) { + if (test_bit(NFS_DELEGATION_REVOKED, &delegation->flags)) { rcu_read_unlock(); nfs_finish_clear_delegation_stateid(state, &stateid); return; } + if (!test_and_clear_bit(NFS_DELEGATION_TEST_EXPIRED, + &delegation->flags)) { + rcu_read_unlock(); + return; + } + cred = get_rpccred(delegation->cred); rcu_read_unlock(); status = nfs41_test_and_free_expired_stateid(server, &stateid, cred); diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 3df0eb52da1c..40a08cd483f0 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1390,6 +1390,8 @@ int nfs4_schedule_stateid_recovery(const struct nfs_server *server, struct nfs4_ if (!nfs4_state_mark_reclaim_nograce(clp, state)) return -EBADF; + nfs_inode_find_delegation_state_and_recover(state->inode, + &state->stateid); dprintk("%s: scheduling stateid recovery for server %s\n", __func__, clp->cl_hostname); nfs4_schedule_state_manager(clp); -- cgit v1.2.1 From 9f0c5124f4a82503ee5d55c60b0b9c6afc3af68b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 5 Sep 2018 14:07:15 -0400 Subject: NFS: Don't open code clearing of delegation state Add a helper for the case when the nfs4 open state has been set to use a delegation stateid, and we want to revert to using the open stateid. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 094c3c09ff00..481787cac4c2 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -1637,6 +1637,14 @@ static void nfs_state_set_delegation(struct nfs4_state *state, write_sequnlock(&state->seqlock); } +static void nfs_state_clear_delegation(struct nfs4_state *state) +{ + write_seqlock(&state->seqlock); + nfs4_stateid_copy(&state->stateid, &state->open_stateid); + clear_bit(NFS_DELEGATED_STATE, &state->flags); + write_sequnlock(&state->seqlock); +} + static int update_open_stateid(struct nfs4_state *state, const nfs4_stateid *open_stateid, const nfs4_stateid *delegation, @@ -2145,10 +2153,7 @@ int nfs4_open_delegation_recall(struct nfs_open_context *ctx, if (IS_ERR(opendata)) return PTR_ERR(opendata); nfs4_stateid_copy(&opendata->o_arg.u.delegation, stateid); - write_seqlock(&state->seqlock); - nfs4_stateid_copy(&state->stateid, &state->open_stateid); - write_sequnlock(&state->seqlock); - clear_bit(NFS_DELEGATED_STATE, &state->flags); + nfs_state_clear_delegation(state); switch (type & (FMODE_READ|FMODE_WRITE)) { case FMODE_READ|FMODE_WRITE: case FMODE_WRITE: @@ -2601,10 +2606,7 @@ static void nfs_finish_clear_delegation_stateid(struct nfs4_state *state, const nfs4_stateid *stateid) { nfs_remove_bad_delegation(state->inode, stateid); - write_seqlock(&state->seqlock); - nfs4_stateid_copy(&state->stateid, &state->open_stateid); - write_sequnlock(&state->seqlock); - clear_bit(NFS_DELEGATED_STATE, &state->flags); + nfs_state_clear_delegation(state); } static void nfs40_clear_delegation_stateid(struct nfs4_state *state) @@ -2672,13 +2674,14 @@ static void nfs41_check_delegation_stateid(struct nfs4_state *state) delegation = rcu_dereference(NFS_I(state->inode)->delegation); if (delegation == NULL) { rcu_read_unlock(); + nfs_state_clear_delegation(state); return; } nfs4_stateid_copy(&stateid, &delegation->stateid); if (test_bit(NFS_DELEGATION_REVOKED, &delegation->flags)) { rcu_read_unlock(); - nfs_finish_clear_delegation_stateid(state, &stateid); + nfs_state_clear_delegation(state); return; } -- cgit v1.2.1