- 11 Feb, 2023 1 commit
-
-
Jeff Layton authored
The nfs4_file table is global, so shutting it down when a containerized nfsd is shut down is wrong and can lead to double-frees. Tear down the nfs4_file_rhltable in nfs4_state_shutdown instead of nfs4_state_shutdown_net. Fixes: d47b295e ("NFSD: Use rhashtable for managing nfs4_file objects") Link: https://bugzilla.redhat.com/show_bug.cgi?id=2169017Reported-by:
JianHong Yin <jiyin@redhat.com> Signed-off-by:
Jeff Layton <jlayton@kernel.org> Signed-off-by:
Chuck Lever <chuck.lever@oracle.com>
-
- 23 Jan, 2023 1 commit
-
-
Jeff Layton authored
nfsd_file_cache_purge is called when the server is shutting down, in which case, tearing things down is generally fine, but it also gets called when the exports cache is flushed. Instead of walking the cache and freeing everything unconditionally, handle it the same as when we have a notification of conflicting access. Fixes: ac3a2585 ("nfsd: rework refcounting in filecache") Reported-by:
Ruben Vestergaard <rubenv@drcmr.dk> Reported-by:
Torkil Svensgaard <torkil@drcmr.dk> Reported-by:
Shachar Kagan <skagan@nvidia.com> Signed-off-by:
-
- 12 Jan, 2023 3 commits
-
-
Dai Ngo authored
Since nfsd4_state_shrinker_count always calls mod_delayed_work with 0 delay, we can replace delayed_work with work_struct to save some space and overhead. Also add the call to cancel_work after unregister the shrinker in nfs4_state_shutdown_net. Signed-off-by:
Dai Ngo <dai.ngo@oracle.com> Reviewed-by:
-
Dai Ngo authored
Currently the nfsd-client shrinker is registered and unregistered at the time the nfsd module is loaded and unloaded. The problem with this is the shrinker is being registered before all of the relevant fields in nfsd_net are initialized when nfsd is started. This can lead to an oops when memory is low and the shrinker is called while nfsd is not running. This patch moves the register/unregister of nfsd-client shrinker from module load/unload time to nfsd startup/shutdown time. Fixes: 44df6f43 ("NFSD: add delegation reaper to react to low memory condition") Reported-by:
Mike Galbraith <efault@gmx.de> Signed-off-by:
-
Xingyuan Mo authored
If signal_pending() returns true, schedule_timeout() will not be executed, causing the waiting task to remain in the wait queue. Fixed by adding a call to finish_wait(), which ensures that the waiting task will always be removed from the wait queue. Fixes: f4e44b39 ("NFSD: delay unmount source's export after inter-server copy completed.") Signed-off-by:
Xingyuan Mo <hdthky0@gmail.com> Reviewed-by:
-
- 07 Jan, 2023 1 commit
-
-
Chuck Lever authored
The premise that "Once an svc thread is scheduled and executing an RPC, no other processes will touch svc_rqst::rq_flags" is false. svc_xprt_enqueue() examines the RQ_BUSY flag in scheduled nfsd threads when determining which thread to wake up next. Fixes: 93155647 ("NFSD: Use only RQ_DROPME to signal the need to drop a reply") Signed-off-by:
-
- 06 Jan, 2023 2 commits
-
-
Chuck Lever authored
The premise that "Once an svc thread is scheduled and executing an RPC, no other processes will touch svc_rqst::rq_flags" is false. svc_xprt_enqueue() examines the RQ_BUSY flag in scheduled nfsd threads when determining which thread to wake up next. Found via KCSAN. Fixes: 28df0988 ("SUNRPC: Use RMW bitops in single-threaded hot paths") Reviewed-by:
-
Jeff Layton authored
Commit fb70bf12 ("NFSD: Instantiate a struct file when creating a regular NFSv4 file") added the ability to cache an open fd over a compound. There are a couple of problems with the way this currently works: It's racy, as a newly-created nfsd_file can end up with its PENDING bit cleared while the nf is hashed, and the nf_file pointer is still zeroed out. Other tasks can find it in this state and they expect to see a valid nf_file, and can oops if nf_file is NULL. Also, there is no guarantee that we'll end up creating a new nfsd_file if one is already in the hash. If an extant entry is in the hash with a valid nf_file, nfs4_get_vfs_file will clobber its nf_file pointer with the value of op_file and the old nf_file will leak. Fix both issues by making a new nfsd_file_acquirei_opened variant that takes an optional file pointer. If one is present when this is called, we'll take a new reference to it instead of trying to open the file. If the nfsd_file already has a valid nf_file, we'll just ignore the optional file and pass the nfsd_file back as-is. Also rework the tracepoints a bit to allow for an "opened" variant and don't try to avoid counting acquisitions in the case where we already have a cached open file. Fixes: fb70bf12 ("NFSD: Instantiate a struct file when creating a regular NFSv4 file") Cc: Trond Myklebust <trondmy@hammerspace.com> Reported-by:
Stanislav Saner <ssaner@redhat.com> Reported-and-Tested-by:
-
- 02 Jan, 2023 1 commit
-
-
Jeff Layton authored
If v4 READDIR operation hits a mountpoint and gets back an error, then it will include that entry in the reply and set RDATTR_ERROR for it to the error. That's fine for "normal" exported filesystems, but on the v4root, we need to be more careful to only expose the existence of dentries that lead to exports. If the mountd upcall times out while checking to see whether a mountpoint on the v4root is exported, then we have no recourse other than to fail the whole operation. Cc: Steve Dickson <steved@redhat.com> Link: https://bugzilla.kernel.org/show_bug.cgi?id=216777Reported-by:
JianHong Yin <yin-jianhong@163.com> Signed-off-by:
-
- 22 Dec, 2022 1 commit
-
-
Jeff Layton authored
Currently, we shut down the filecache before trying to clean up the stateids that depend on it. This leads to the kernel trying to free an nfsd_file twice, and a refcount overput on the nf_mark. Change the shutdown procedure to tear down all of the stateids prior to shutting down the filecache. Reported-and-tested-by:
Wang Yugui <wangyugui@e16-tech.com> Signed-off-by:
-
- 14 Dec, 2022 1 commit
-
-
Dai Ngo authored
Problem caused by source's vfsmount being unmounted but remains on the delayed unmount list. This happens when nfs42_ssc_open() return errors. Fixed by removing nfsd4_interssc_connect(), leave the vfsmount for the laundromat to unmount when idle time expires. We don't need to call nfs_do_sb_deactive when nfs42_ssc_open return errors since the file was not opened so nfs_server->active was not incremented. Same as in nfsd4_copy, if we fail to launch nfsd4_do_async_copy thread then there's no need to call nfs_do_sb_deactive Reported-by:
-
- 12 Dec, 2022 1 commit
-
-
Dan Aloni authored
On error situation `clp->cl_cb_conn.cb_xprt` should not be given a reference to the xprt otherwise both client cleanup and the error handling path of the caller call to put it. Better to delay handing over the reference to a later branch. [ 72.530665] refcount_t: underflow; use-after-free. [ 72.531933] WARNING: CPU: 0 PID: 173 at lib/refcount.c:28 refcount_warn_saturate+0xcf/0x120 [ 72.533075] Modules linked in: nfsd(OE) nfsv4(OE) nfsv3(OE) nfs(OE) lockd(OE) compat_nfs_ssc(OE) nfs_acl(OE) rpcsec_gss_krb5(OE) auth_rpcgss(OE) rpcrdma(OE) dns_resolver fscache netfs grace rdma_cm iw_cm ib_cm sunrpc(OE) mlx5_ib mlx5_core mlxfw pci_hyperv_intf ib_uverbs ib_core xt_MASQUERADE nf_conntrack_netlink nft_counter xt_addrtype nft_compat br_netfilter bridge stp llc nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 ip_set overlay nf_tables nfnetlink crct10dif_pclmul crc32_pclmul ghash_clmulni_intel xfs serio_raw virtio_net virtio_blk net_failover failover fuse [last unloaded: sunrpc] [ 72.540389] CPU: 0 PID: 173 Comm: kworker/u16:5 Tainted: G OE 5.15.82-dan #1 [ 72.541511] Hardware name: Red Hat KVM/RHEL-AV, BIOS 1.16.0-3.module+el8.7.0+1084+97b81f61 04/01/2014 [ 72.542717] Workqueue: nfsd4_callbacks nfsd4_run_cb_work [nfsd] [ 72.543575] RIP: 0010:refcount_warn_saturate+0xcf/0x120 [ 72.544299] Code: 55 00 0f 0b 5d e9 01 50 98 00 80 3d 75 9e 39 08 00 0f 85 74 ff ff ff 48 c7 c7 e8 d1 60 8e c6 05 61 9e 39 08 01 e8 f6 51 55 00 <0f> 0b 5d e9 d9 4f 98 00 80 3d 4b 9e 39 08 00 0f 85 4c ff ff ff 48 [ 72.546666] RSP: 0018:ffffb3f841157cf0 EFLAGS: 00010286 [ 72.547393] RAX: 0000000000000026 RBX: ffff89ac6231d478 RCX: 0000000000000000 [ 72.548324] RDX: ffff89adb7c2c2c0 RSI: ffff89adb7c205c0 RDI: ffff89adb7c205c0 [ 72.549271] RBP: ffffb3f841157cf0 R08: 0000000000000000 R09: c0000000ffefffff [ 72.550209] R10: 0000000000000001 R11: ffffb3f841157ad0 R12: ffff89ac6231d180 [ 72.551142] R13: ffff89ac6231d478 R14: ffff89ac40c06180 R15: ffff89ac6231d4b0 [ 72.552089] FS: 0000000000000000(0000) GS:ffff89adb7c00000(0000) knlGS:0000000000000000 [ 72.553175] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 72.553934] CR2: 0000563a310506a8 CR3: 0000000109a66000 CR4: 0000000000350ef0 [ 72.554874] Call Trace: [ 72.555278] <TASK> [ 72.555614] svc_xprt_put+0xaf/0xe0 [sunrpc] [ 72.556276] nfsd4_process_cb_update.isra.11+0xb7/0x410 [nfsd] [ 72.557087] ? update_load_avg+0x82/0x610 [ 72.557652] ? cpuacct_charge+0x60/0x70 [ 72.558212] ? dequeue_entity+0xdb/0x3e0 [ 72.558765] ? queued_spin_unlock+0x9/0x20 [ 72.559358] nfsd4_run_cb_work+0xfc/0x270 [nfsd] [ 72.560031] process_one_work+0x1df/0x390 [ 72.560600] worker_thread+0x37/0x3b0 [ 72.561644] ? process_one_work+0x390/0x390 [ 72.562247] kthread+0x12f/0x150 [ 72.562710] ? set_kthread_struct+0x50/0x50 [ 72.563309] ret_from_fork+0x22/0x30 [ 72.563818] </TASK> [ 72.564189] ---[ end trace 031117b1c72ec616 ]--- [ 72.566019] list_add corruption. next->prev should be prev (ffff89ac4977e538), but was ffff89ac4763e018. (next=ffff89ac4763e018). [ 72.567647] ------------[ cut here ]------------ Fixes: a4abc6b1 ("nfsd: Fix svc_xprt refcnt leak when setup callback client failed") Cc: Xiyu Yang <xiyuyang19@fudan.edu.cn> Cc: J. Bruce Fields <bfields@redhat.com> Signed-off-by:
Dan Aloni <dan.aloni@vastdata.com> Reviewed-by:
-
- 11 Dec, 2022 1 commit
-
-
Jeff Layton authored
The filecache refcounting is a bit non-standard for something searchable by RCU, in that we maintain a sentinel reference while it's hashed. This in turn requires that we have to do things differently in the "put" depending on whether its hashed, which we believe to have led to races. There are other problems in here too. nfsd_file_close_inode_sync can end up freeing an nfsd_file while there are still outstanding references to it, and there are a number of subtle ToC/ToU races. Rework the code so that the refcount is what drives the lifecycle. When the refcount goes to zero, then unhash and rcu free the object. A task searching for a nfsd_file is allowed to bump its refcount, but only if it's not already 0. Ensure that we don't make any other changes to it until a reference is held. With this change, the LRU carries a reference. Take special care to deal with it when removing an entry from the list, and ensure that we only repurpose the nf_lru list_head when the refcount is 0 to ensure exclusive access to it. Signed-off-by:
-
- 10 Dec, 2022 21 commits
-
-
Kees Cook authored
When built with Control Flow Integrity, function prototypes between caller and function declaration must match. These mismatches are visible at compile time with the new -Wcast-function-type-strict in Clang[1]. There were 97 warnings produced by NFS. For example: fs/nfsd/nfs4xdr.c:2228:17: warning: cast from '__be32 (*)(struct nfsd4_compoundargs *, struct nfsd4_access *)' (aka 'unsigned int (*)(struct nfsd4_compoundargs *, struct nfsd4_access *)') to 'nfsd4_dec' (aka 'unsigned int (*)(struct nfsd4_compoundargs *, void *)') converts to incompatible function type [-Wcast-function-type-strict] [OP_ACCESS] = (nfsd4_dec)nfsd4_decode_access, ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The enc/dec callbacks were defined as passing "void *" as the second argument, but were being implicitly cast to a new type. Replace the argument with union nfsd4_op_u, and perform explicit member selection in the function body. There are no resulting binary differences. Changes were made mechanically using the following Coccinelle script, with minor by-hand fixes for members that didn't already match their existing argument name: @find@ identifier func; type T, opsT; identifier ops, N; @@ opsT ops[] = { [N] = (T) func, }; @already_void@ identifier find.func; identifier name; @@ func(..., -void +union nfsd4_op_u *name) { ... } @proto depends on !already_void@ identifier find.func; type T; identifier name; position p; @@ func@p(..., T name ) { ... } @script:python get_member@ type_name << proto.T; member; @@ coccinelle.member = cocci.make_ident(type_name.split("_", 1)[1].split(' ',1)[0]) @convert@ identifier find.func; type proto.T; identifier proto.name; position proto.p; identifier get_member.member; @@ func@p(..., - T name + union nfsd4_op_u *u ) { + T name = &u->member; ... } @cast@ identifier find.func; type T, opsT; identifier ops, N; @@ opsT ops[] = { [N] = - (T) func, }; Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Jeff Layton <jlayton@kernel.org> Cc: Gustavo A. R. Silva <gustavoars@kernel.org> Cc: linux-nfs@vger.kernel.org Signed-off-by:Kees Cook <keescook@chromium.org> Signed-off-by:
-
Chuck Lever authored
If a zero length is passed to kmalloc() it returns 0x10, which is not a valid address. gss_verify_mic() subsequently crashes when it attempts to dereference that pointer. Instead of allocating this memory on every call based on an untrusted size value, use a piece of dynamically-allocated scratch memory that is always available. Signed-off-by:
-
Chuck Lever authored
Clean up: Simplify the tracepoint's only call site. Also, I noticed that when svc_authenticate() returns SVC_COMPLETE, it leaves rq_auth_stat set to an error value. That doesn't need to be recorded in the trace log. Signed-off-by:
-
Chuck Lever authored
Clean up: NFSv2 has the only two usages of rpc_drop_reply in the NFSD code base. Since NFSv2 is going away at some point, replace these in order to simplify the "drop this reply?" check in nfsd_dispatch(). Signed-off-by:
-
Chuck Lever authored
Make it more evident how xdr_write_pages() updates the tail buffer by using the convention of naming the iov pointer variable "tail". I spent more than a couple of hours chasing through code to understand this, so someone is likely to find this useful later. Signed-off-by:
-
Chuck Lever authored
Fixes: 030d794b ("SUNRPC: Use gssproxy upcall for server RPCGSS authentication.") Signed-off-by:
-
Dai Ngo authored
Add tracepoints to trace start and end of CB_RECALL_ANY operation. Signed-off-by:
-
Dai Ngo authored
The delegation reaper is called by nfsd memory shrinker's on the 'count' callback. It scans the client list and sends the courtesy CB_RECALL_ANY to the clients that hold delegations. To avoid flooding the clients with CB_RECALL_ANY requests, the delegation reaper sends only one CB_RECALL_ANY request to each client per 5 seconds. Signed-off-by:
-
Dai Ngo authored
Add XDR encode and decode function for CB_RECALL_ANY. Signed-off-by:
-
Dai Ngo authored
Refactoring courtesy_client_reaper to generic low memory shrinker so it can be used for other purposes. Signed-off-by:
-
Chuck Lever authored
Steven Rostedt says: > The include/trace/events/ directory should only hold files that > are to create events, not headers that hold helper functions. > > Can you please move them out of include/trace/events/ as that > directory is "special" in the creation of events. Signed-off-by:
Leon Romanovsky <leonro@nvidia.com> Acked-by:
Steven Rostedt (Google) <rostedt@goodmis.org> Acked-by:
Anna Schumaker <Anna.Schumaker@Netapp.com>
-
Brian Foster authored
_nfsd_copy_file_range() calls vfs_fsync_range() with an offset and count (bytes written), but the former wants the start and end bytes of the range to sync. Fix it up. Fixes: eac0b17a ("NFSD add vfs_fsync after async copy is done") Signed-off-by:
Brian Foster <bfoster@redhat.com> Tested-by:
-
Jeff Layton authored
We currently do a lock_to_openmode call based on the arguments from the NLM_UNLOCK call, but that will always set the fl_type of the lock to F_UNLCK, and the O_RDONLY descriptor is always chosen. Fix it to use the file_lock from the block instead. Signed-off-by:
-
Jeff Layton authored
Shared locks are set on O_RDONLY descriptors and exclusive locks are set on O_WRONLY ones. nlmsvc_unlock however calls vfs_lock_file twice, once for each descriptor, but it doesn't reset fl_file. Ensure that it does. Signed-off-by:
-
Jeff Layton authored
Signed-off-by:
-
Xiu Jianfeng authored
Use struct_size() helper to simplify the code, no functional changes. Signed-off-by:
Xiu Jianfeng <xiujianfeng@huawei.com> Reviewed-by:
-
Jeff Layton authored
With the addition of POSIX ACLs to struct nfsd_attrs, we no longer return an error if setting the ACL fails. Ensure we return the na_aclerr error on SETATTR if there is one. Fixes: c0cbe707 ("NFSD: add posix ACLs to struct nfsd_attrs") Cc: Neil Brown <neilb@suse.de> Reported-by:
Yongcheng Yang <yoyang@redhat.com> Signed-off-by:
-
Trond Myklebust authored
vfs_lock_file() expects the struct file_lock to be fully initialised by the caller. Re-exported NFSv3 has been seen to Oops if the fl_file field is NULL. Fixes: aec15824 ("lockd: set fl_owner when unlocking files") Signed-off-by:
Trond Myklebust <trond.myklebust@hammerspace.com> Reviewed-by:
-
Chuck Lever authored
Add a tracepoint to capture the number of filecache-triggered fsync calls and which files needed it. Also, record when an fsync triggers a write verifier reset. Examples: <...>-97 [007] 262.505611: nfsd_file_free: inode=0xffff888171e08140 ref=0 flags=GC may=WRITE nf_file=0xffff8881373d2400 <...>-97 [007] 262.505612: nfsd_file_fsync: inode=0xffff888171e08140 ref=0 flags=GC may=WRITE nf_file=0xffff8881373d2400 ret=0 <...>-97 [007] 262.505623: nfsd_file_free: inode=0xffff888171e08dc0 ref=0 flags=GC may=WRITE nf_file=0xffff8881373d1e00 <...>-97 [007] 262.505624: nfsd_file_fsync: inode=0xffff888171e08dc0 ref=0 flags=GC may=WRITE nf_file=0xffff8881373d1e00 ret=0 Signed-off-by:
-
Li zeming authored
The svc_ungetu32 function is not used, you could remove it. Signed-off-by:
Li zeming <zeming@nfschina.com> Signed-off-by:
-
Jeff Layton authored
We don't really care whether there are hashed entries when it comes to scheduling the laundrette. They might all be non-gc entries, after all. We only want to schedule it if there are entries on the LRU. Switch to using list_lru_count, and move the check into nfsd_file_gc_worker. The other callsite in nfsd_file_put doesn't need to count entries, since it only schedules the laundrette after adding an entry to the LRU. Signed-off-by:
-
- 28 Nov, 2022 6 commits
-
-
Jeff Layton authored
In a coming patch, we're going to rework how the filecache refcounting works. Move some code around in the function to reduce the churn in the later patches, and rename some of the functions with (hopefully) clearer names: nfsd_file_flush becomes nfsd_file_fsync, and nfsd_file_unhash_and_dispose is renamed to nfsd_file_unhash_and_queue. Also, the nfsd_file_put_final tracepoint is renamed to nfsd_file_free, to better match the name of the function from which it's called. Signed-off-by:
NeilBrown <neilb@suse.de> Signed-off-by:
-
Jeff Layton authored
We're counting mapping->nrpages, but not all of those are necessarily dirty. We don't really have a simple way to count just the dirty pages, so just remove this stat since it's not accurate. Signed-off-by:
-
Chuck Lever authored
We recently received a patch for fs/exportfs/expfs.c, but there isn't a subsystem maintainer listed for fs/exportfs: Christian Brauner <brauner@kernel.org> (commit_signer:2/2=100%,authored:1/2=50%,added_lines:3/6=50%,removed_lines:2/6=33%) Al Viro <viro@zeniv.linux.org.uk> (commit_signer:1/2=50%,authored:1/2=50%,added_lines:3/6=50%,removed_lines:4/6=67%) Miklos Szeredi <mszeredi@redhat.com> (commit_signer:1/2=50%) Amir Goldstein <amir73il@gmail.com> (commit_signer:1/2=50%) linux-kernel@vger.kernel.org (open list) Neil says: > Looking at recent commits, patches come in through multiple > different trees. > nfsd certainly has an interest in expfs.c. The only other user is > name_to_handle/open_by_handle API. > I see it as primarily nfsd functionality which is useful enough to > be exported directly to user-space. > (It was created by me when I was nfsd maintainer - does that > count?) Suggested-by:
-
Chuck Lever authored
Add a missing SPDX header. Signed-off-by:
-
Chuck Lever authored
fh_match() is costly, especially when filehandles are large (as is the case for NFSv4). It needs to be used sparingly when searching data structures. Unfortunately, with common workloads, I see multiple thousands of objects stored in file_hashtbl[], which has just 256 buckets, making its bucket hash chains quite lengthy. Walking long hash chains with the state_lock held blocks other activity that needs that lock. Sizable hash chains are a common occurrance once the server has handed out some delegations, for example -- IIUC, each delegated file is held open on the server by an nfs4_file object. To help mitigate the cost of searching with fh_match(), replace the nfs4_file hash table with an rhashtable, which can dynamically resize its bucket array to minimize hash chain length. The result of this modification is an improvement in the latency of NFSv4 operations, and the reduction of nfsd CPU utilization due to eliminating the cost of multiple calls to fh_match() and reducing the CPU cache misses incurred while walking long hash chains in the nfs4_file hash table. Signed-off-by:
-
Chuck Lever authored
find_file() is now the only caller of find_file_locked(), so just fold these two together. Signed-off-by:
-
