From fcf857ee1958e9247298251f7615d0c76f1e9b38 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 1 Feb 2025 14:59:02 -0500 Subject: [PATCH 1/7] NFS: O_DIRECT writes must check and adjust the file length While it is uncommon for delegations to be held while O_DIRECT writes are in progress, it is possible. The xfstests generic/647 and generic/729 both end up triggering that state, and end up failing due to the fact that the file size is not adjusted. Reported-by: Chuck Lever Link: https://bugzilla.kernel.org/show_bug.cgi?id=219738 Cc: stable@vger.kernel.org Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/direct.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index f45beea92d03..40e13c9a2873 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -130,6 +130,20 @@ static void nfs_direct_truncate_request(struct nfs_direct_req *dreq, dreq->count = req_start; } +static void nfs_direct_file_adjust_size_locked(struct inode *inode, + loff_t offset, size_t count) +{ + loff_t newsize = offset + (loff_t)count; + loff_t oldsize = i_size_read(inode); + + if (newsize > oldsize) { + i_size_write(inode, newsize); + NFS_I(inode)->cache_validity &= ~NFS_INO_INVALID_SIZE; + trace_nfs_size_grow(inode, newsize); + nfs_inc_stats(inode, NFSIOS_EXTENDWRITE); + } +} + /** * nfs_swap_rw - NFS address space operation for swap I/O * @iocb: target I/O control block @@ -741,6 +755,7 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr) struct nfs_direct_req *dreq = hdr->dreq; struct nfs_commit_info cinfo; struct nfs_page *req = nfs_list_entry(hdr->pages.next); + struct inode *inode = dreq->inode; int flags = NFS_ODIRECT_DONE; trace_nfs_direct_write_completion(dreq); @@ -762,6 +777,10 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr) } spin_unlock(&dreq->lock); + spin_lock(&inode->i_lock); + nfs_direct_file_adjust_size_locked(inode, dreq->io_start, dreq->count); + spin_unlock(&inode->i_lock); + while (!list_empty(&hdr->pages)) { req = nfs_list_entry(hdr->pages.next); From 88025c67fe3c025a0123bc7af50535b97f7af89a Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 1 Feb 2025 14:59:03 -0500 Subject: [PATCH 2/7] NFS: Adjust delegated timestamps for O_DIRECT reads and writes Adjust the timestamps if O_DIRECT is being combined with attribute delegations. Fixes: e12912d94137 ("NFSv4: Add support for delegated atime and mtime attributes") Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/direct.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 40e13c9a2873..f32f8d7c9122 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -56,6 +56,7 @@ #include #include +#include "delegation.h" #include "internal.h" #include "iostat.h" #include "pnfs.h" @@ -286,6 +287,8 @@ static void nfs_direct_read_completion(struct nfs_pgio_header *hdr) nfs_direct_count_bytes(dreq, hdr); spin_unlock(&dreq->lock); + nfs_update_delegated_atime(dreq->inode); + while (!list_empty(&hdr->pages)) { struct nfs_page *req = nfs_list_entry(hdr->pages.next); struct page *page = req->wb_page; @@ -779,6 +782,7 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr) spin_lock(&inode->i_lock); nfs_direct_file_adjust_size_locked(inode, dreq->io_start, dreq->count); + nfs_update_delegated_mtime_locked(dreq->inode); spin_unlock(&inode->i_lock); while (!list_empty(&hdr->pages)) { From 5bbd6e863b15a85221e49b9bdb2d5d8f0bb91f3d Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 1 Feb 2025 15:00:02 -0500 Subject: [PATCH 3/7] SUNRPC: Prevent looping due to rpc_signal_task() races If rpc_signal_task() is called while a task is in an rpc_call_done() callback function, and the latter calls rpc_restart_call(), the task can end up looping due to the RPC_TASK_SIGNALLED flag being set without the tk_rpc_status being set. Removing the redundant mechanism for signalling the task fixes the looping behaviour. Reported-by: Li Lingfeng Fixes: 39494194f93b ("SUNRPC: Fix races with rpc_killall_tasks()") Signed-off-by: Trond Myklebust Reviewed-by: Jeff Layton Signed-off-by: Anna Schumaker --- include/linux/sunrpc/sched.h | 3 +-- include/trace/events/sunrpc.h | 3 +-- net/sunrpc/sched.c | 2 -- 3 files changed, 2 insertions(+), 6 deletions(-) diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h index fec1e8a1570c..eac57914dcf3 100644 --- a/include/linux/sunrpc/sched.h +++ b/include/linux/sunrpc/sched.h @@ -158,7 +158,6 @@ enum { RPC_TASK_NEED_XMIT, RPC_TASK_NEED_RECV, RPC_TASK_MSG_PIN_WAIT, - RPC_TASK_SIGNALLED, }; #define rpc_test_and_set_running(t) \ @@ -171,7 +170,7 @@ enum { #define RPC_IS_ACTIVATED(t) test_bit(RPC_TASK_ACTIVE, &(t)->tk_runstate) -#define RPC_SIGNALLED(t) test_bit(RPC_TASK_SIGNALLED, &(t)->tk_runstate) +#define RPC_SIGNALLED(t) (READ_ONCE(task->tk_rpc_status) == -ERESTARTSYS) /* * Task priorities. diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h index b13dc275ef4a..851841336ee6 100644 --- a/include/trace/events/sunrpc.h +++ b/include/trace/events/sunrpc.h @@ -360,8 +360,7 @@ TRACE_EVENT(rpc_request, { (1UL << RPC_TASK_ACTIVE), "ACTIVE" }, \ { (1UL << RPC_TASK_NEED_XMIT), "NEED_XMIT" }, \ { (1UL << RPC_TASK_NEED_RECV), "NEED_RECV" }, \ - { (1UL << RPC_TASK_MSG_PIN_WAIT), "MSG_PIN_WAIT" }, \ - { (1UL << RPC_TASK_SIGNALLED), "SIGNALLED" }) + { (1UL << RPC_TASK_MSG_PIN_WAIT), "MSG_PIN_WAIT" }) DECLARE_EVENT_CLASS(rpc_task_running, diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index cef623ea1506..9b45fbdc90ca 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -864,8 +864,6 @@ void rpc_signal_task(struct rpc_task *task) if (!rpc_task_set_rpc_status(task, -ERESTARTSYS)) return; trace_rpc_task_signalled(task, task->tk_action); - set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate); - smp_mb__after_atomic(); queue = READ_ONCE(task->tk_waitqueue); if (queue) rpc_wake_up_queued_task(queue, task); From 8f8df955f078e1a023ee55161935000a67651f38 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 1 Feb 2025 15:00:09 -0500 Subject: [PATCH 4/7] NFSv4: Fix a deadlock when recovering state on a sillyrenamed file If the file is sillyrenamed, and slated for delete on close, it is possible for a server reboot to triggeer an open reclaim, with can again race with the application call to close(). When that happens, the call to put_nfs_open_context() can trigger a synchronous delegreturn call which deadlocks because it is not marked as privileged. Instead, ensure that the call to nfs4_inode_return_delegation_on_close() catches the delegreturn, and schedules it asynchronously. Reported-by: Li Lingfeng Fixes: adb4b42d19ae ("Return the delegation when deleting sillyrenamed files") Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/delegation.c | 37 +++++++++++++++++++++++++++++++++++++ fs/nfs/delegation.h | 1 + fs/nfs/nfs4proc.c | 3 +++ 3 files changed, 41 insertions(+) diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index 035ba52742a5..4db912f56230 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -780,6 +780,43 @@ int nfs4_inode_return_delegation(struct inode *inode) return 0; } +/** + * nfs4_inode_set_return_delegation_on_close - asynchronously return a delegation + * @inode: inode to process + * + * This routine is called to request that the delegation be returned as soon + * as the file is closed. If the file is already closed, the delegation is + * immediately returned. + */ +void nfs4_inode_set_return_delegation_on_close(struct inode *inode) +{ + struct nfs_delegation *delegation; + struct nfs_delegation *ret = NULL; + + if (!inode) + return; + rcu_read_lock(); + delegation = nfs4_get_valid_delegation(inode); + if (!delegation) + goto out; + spin_lock(&delegation->lock); + if (!delegation->inode) + goto out_unlock; + if (list_empty(&NFS_I(inode)->open_files) && + !test_and_set_bit(NFS_DELEGATION_RETURNING, &delegation->flags)) { + /* Refcount matched in nfs_end_delegation_return() */ + ret = nfs_get_delegation(delegation); + } else + set_bit(NFS_DELEGATION_RETURN_IF_CLOSED, &delegation->flags); +out_unlock: + spin_unlock(&delegation->lock); + if (ret) + nfs_clear_verifier_delegated(inode); +out: + rcu_read_unlock(); + nfs_end_delegation_return(inode, ret, 0); +} + /** * nfs4_inode_return_delegation_on_close - asynchronously return a delegation * @inode: inode to process diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h index 71524d34ed20..8ff5ab9c5c25 100644 --- a/fs/nfs/delegation.h +++ b/fs/nfs/delegation.h @@ -49,6 +49,7 @@ void nfs_inode_reclaim_delegation(struct inode *inode, const struct cred *cred, unsigned long pagemod_limit, u32 deleg_type); int nfs4_inode_return_delegation(struct inode *inode); void nfs4_inode_return_delegation_on_close(struct inode *inode); +void nfs4_inode_set_return_delegation_on_close(struct inode *inode); int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *stateid); void nfs_inode_evict_delegation(struct inode *inode); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index df9669d4ded7..c25ecdb76d30 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3906,8 +3906,11 @@ nfs4_atomic_open(struct inode *dir, struct nfs_open_context *ctx, static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync) { + struct dentry *dentry = ctx->dentry; if (ctx->state == NULL) return; + if (dentry->d_flags & DCACHE_NFSFS_RENAMED) + nfs4_inode_set_return_delegation_on_close(d_inode(dentry)); if (is_sync) nfs4_close_sync(ctx->state, _nfs4_ctx_to_openmode(ctx)); else From 7a2f6f7687c5f7083a35317cddec5ad9fa491443 Mon Sep 17 00:00:00 2001 From: Benjamin Coddington Date: Tue, 11 Feb 2025 12:31:57 -0500 Subject: [PATCH 5/7] SUNRPC: Handle -ETIMEDOUT return from tlshd If the TLS handshake attempt returns -ETIMEDOUT, we currently translate that error into -EACCES. This becomes problematic for cases where the RPC layer is attempting to re-connect in paths that don't resonably handle -EACCES, for example: writeback. The RPC layer can handle -ETIMEDOUT quite well, however - so if the handshake returns this error let's just pass it along. Fixes: 75eb6af7acdf ("SUNRPC: Add a TCP-with-TLS RPC transport class") Signed-off-by: Benjamin Coddington Signed-off-by: Anna Schumaker --- net/sunrpc/xprtsock.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index c60936d8cef7..6b80b2aaf763 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -2581,7 +2581,15 @@ static void xs_tls_handshake_done(void *data, int status, key_serial_t peerid) struct sock_xprt *lower_transport = container_of(lower_xprt, struct sock_xprt, xprt); - lower_transport->xprt_err = status ? -EACCES : 0; + switch (status) { + case 0: + case -EACCES: + case -ETIMEDOUT: + lower_transport->xprt_err = status; + break; + default: + lower_transport->xprt_err = -EACCES; + } complete(&lower_transport->handshake_done); xprt_put(lower_xprt); } From 1f7a4f98c11fbeb18ed21f3b3a497e90a50ad2e0 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 25 Feb 2025 15:52:21 +0100 Subject: [PATCH 6/7] sunrpc: suppress warnings for unused procfs functions There is a warning about unused variables when building with W=1 and no procfs: net/sunrpc/cache.c:1660:30: error: 'cache_flush_proc_ops' defined but not used [-Werror=unused-const-variable=] 1660 | static const struct proc_ops cache_flush_proc_ops = { | ^~~~~~~~~~~~~~~~~~~~ net/sunrpc/cache.c:1622:30: error: 'content_proc_ops' defined but not used [-Werror=unused-const-variable=] 1622 | static const struct proc_ops content_proc_ops = { | ^~~~~~~~~~~~~~~~ net/sunrpc/cache.c:1598:30: error: 'cache_channel_proc_ops' defined but not used [-Werror=unused-const-variable=] 1598 | static const struct proc_ops cache_channel_proc_ops = { | ^~~~~~~~~~~~~~~~~~~~~~ These are used inside of an #ifdef, so replacing that with an IS_ENABLED() check lets the compiler see how they are used while still dropping them during dead code elimination. Fixes: dbf847ecb631 ("knfsd: allow cache_register to return error on failure") Reviewed-by: Jeff Layton Acked-by: Chuck Lever Signed-off-by: Arnd Bergmann Signed-off-by: Anna Schumaker --- net/sunrpc/cache.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index cb279eb9ac4b..7ce5e28a6c03 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -1674,12 +1674,14 @@ static void remove_cache_proc_entries(struct cache_detail *cd) } } -#ifdef CONFIG_PROC_FS static int create_cache_proc_entries(struct cache_detail *cd, struct net *net) { struct proc_dir_entry *p; struct sunrpc_net *sn; + if (!IS_ENABLED(CONFIG_PROC_FS)) + return 0; + sn = net_generic(net, sunrpc_net_id); cd->procfs = proc_mkdir(cd->name, sn->proc_net_rpc); if (cd->procfs == NULL) @@ -1707,12 +1709,6 @@ out_nomem: remove_cache_proc_entries(cd); return -ENOMEM; } -#else /* CONFIG_PROC_FS */ -static int create_cache_proc_entries(struct cache_detail *cd, struct net *net) -{ - return 0; -} -#endif void __init cache_initialize(void) { From 9084ed79ddaaaa1ec01cd304af9fb532c26252db Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Thu, 20 Feb 2025 14:29:36 -0500 Subject: [PATCH 7/7] lsm,nfs: fix memory leak of lsm_context commit b530104f50e8 ("lsm: lsm_context in security_dentry_init_security") did not preserve the lsm id for subsequent release calls, which results in a memory leak. Fix it by saving the lsm id in the nfs4_label and providing it on the subsequent release call. Fixes: b530104f50e8 ("lsm: lsm_context in security_dentry_init_security") Signed-off-by: Stephen Smalley Acked-by: Paul Moore Acked-by: Casey Schaufler Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 7 ++++--- include/linux/nfs4.h | 1 + 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index c25ecdb76d30..6e95db6c17e9 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -133,6 +133,7 @@ nfs4_label_init_security(struct inode *dir, struct dentry *dentry, if (err) return NULL; + label->lsmid = shim.id; label->label = shim.context; label->len = shim.len; return label; @@ -145,7 +146,7 @@ nfs4_label_release_security(struct nfs4_label *label) if (label) { shim.context = label->label; shim.len = label->len; - shim.id = LSM_ID_UNDEF; + shim.id = label->lsmid; security_release_secctx(&shim); } } @@ -6272,7 +6273,7 @@ static int _nfs4_get_security_label(struct inode *inode, void *buf, size_t buflen) { struct nfs_server *server = NFS_SERVER(inode); - struct nfs4_label label = {0, 0, buflen, buf}; + struct nfs4_label label = {0, 0, 0, buflen, buf}; u32 bitmask[3] = { 0, 0, FATTR4_WORD2_SECURITY_LABEL }; struct nfs_fattr fattr = { @@ -6377,7 +6378,7 @@ static int nfs4_do_set_security_label(struct inode *inode, static int nfs4_set_security_label(struct inode *inode, const void *buf, size_t buflen) { - struct nfs4_label ilabel = {0, 0, buflen, (char *)buf }; + struct nfs4_label ilabel = {0, 0, 0, buflen, (char *)buf }; struct nfs_fattr *fattr; int status; diff --git a/include/linux/nfs4.h b/include/linux/nfs4.h index 71fbebfa43c7..9ac83ca88326 100644 --- a/include/linux/nfs4.h +++ b/include/linux/nfs4.h @@ -47,6 +47,7 @@ struct nfs4_acl { struct nfs4_label { uint32_t lfs; uint32_t pi; + u32 lsmid; u32 len; char *label; };