linux.git/fs/nfsd, branch v6.14

Merge tag 'nfsd-6.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux

2025-02-10T21:11:24+00:00

Pull nfsd fixes from Chuck Lever:
 "Fixes for new bugs:
   - A fix for CB_GETATTR reply decoding was not quite correct
   - Fix the NFSD connection limiting logic
   - Fix a bug in the new session table resizing logic

  Bugs that pre-date v6.14:
   - Support for courteous clients (5.19) introduced a shutdown hang
   - Fix a crash in the filecache laundrette (6.9)
   - Fix a zero-day crash in NFSD's NFSv3 ACL implementation"

* tag 'nfsd-6.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux:
  NFSD: Fix CB_GETATTR status fix
  NFSD: fix hang in nfsd4_shutdown_callback
  nfsd: fix __fh_verify for localio
  nfsd: fix uninitialised slot info when a request is retried
  nfsd: validate the nfsd_serv pointer before calling svc_wake_up
  nfsd: clear acl_access/acl_default after releasing them

NFSD: Fix CB_GETATTR status fix

2025-02-10T18:31:28+00:00

Jeff says:

Now that I look, 1b3e26a5ccbf is wrong. The patch on the ml was correct, but
the one that got committed is different. It should be:

    status = decode_cb_op_status(xdr, OP_CB_GETATTR, &cb->cb_status);
    if (unlikely(status || cb->cb_status))

If "status" is non-zero, decoding failed (usu. BADXDR), but we also want to
bail out and not decode the rest of the call if the decoded cb_status is
non-zero. That's not happening here, cb_seq_status has already been checked and
is non-zero, so this ends up trying to decode the rest of the CB_GETATTR reply
when it doesn't exist.

Reported-by: Jeff Layton 
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219737
Fixes: 1b3e26a5ccbf ("NFSD: fix decoding in nfs4_xdr_dec_cb_getattr")
Reviewed-by: Jeff Layton 
Signed-off-by: Chuck Lever

NFSD: fix hang in nfsd4_shutdown_callback

2025-02-10T18:31:16+00:00

If nfs4_client is in courtesy state then there is no point to send
the callback. This causes nfsd4_shutdown_callback to hang since
cl_cb_inflight is not 0. This hang lasts about 15 minutes until TCP
notifies NFSD that the connection was dropped.

This patch modifies nfsd4_run_cb_work to skip the RPC call if
nfs4_client is in courtesy state.

Signed-off-by: Dai Ngo 
Fixes: 66af25799940 ("NFSD: add courteous server support for thread with only delegation")
Cc: stable@vger.kernel.org
Reviewed-by: Jeff Layton 
Signed-off-by: Chuck Lever

nfsd: fix __fh_verify for localio

2025-02-10T18:31:11+00:00

__fh_verify() added a call to svc_xprt_set_valid() to help do connection
management but during LOCALIO path rqstp argument is NULL, leading to
NULL pointer dereferencing and a crash.

Fixes: eccbbc7c00a5 ("nfsd: don't use sv_nrthreads in connection limiting calculations.")
Signed-off-by: Olga Kornievskaia 
Reviewed-by: Jeff Layton 
Signed-off-by: Chuck Lever

nfsd: fix uninitialised slot info when a request is retried

2025-02-10T18:30:55+00:00

A recent patch moved the assignment of seq->maxslots from before the
test for a resent request (which ends with a goto) to after, resulting
in it not being run in that case.  This results in the server returning
bogus "high slot id" and "target high slot id" values.

The assignments to ->maxslots and ->target_maxslots need to be *after*
the out: label so that the correct values are returned in replies to
requests that are served from cache.

Fixes: 60aa6564317d ("nfsd: allocate new session-based DRC slots on demand.")
Signed-off-by: NeilBrown 
Reviewed-by: Jeff Layton 
Signed-off-by: Chuck Lever

nfsd: validate the nfsd_serv pointer before calling svc_wake_up

2025-02-02T16:31:58+00:00

nfsd_file_dispose_list_delayed can be called from the filecache
laundrette, which is shut down after the nfsd threads are shut down and
the nfsd_serv pointer is cleared. If nn->nfsd_serv is NULL then there
are no threads to wake.

Ensure that the nn->nfsd_serv pointer is non-NULL before calling
svc_wake_up in nfsd_file_dispose_list_delayed. This is safe since the
svc_serv is not freed until after the filecache laundrette is cancelled.

Reported-by: Salvatore Bonaccorso 
Closes: https://bugs.debian.org/1093734
Fixes: ffb402596147 ("nfsd: Don't leave work of closing files to a work queue")
Cc: stable@vger.kernel.org
Signed-off-by: Jeff Layton 
Reviewed-by: NeilBrown 
Signed-off-by: Chuck Lever

nfsd: clear acl_access/acl_default after releasing them

2025-02-02T16:31:45+00:00

If getting acl_default fails, acl_access and acl_default will be released
simultaneously. However, acl_access will still retain a pointer pointing
to the released posix_acl, which will trigger a WARNING in
nfs3svc_release_getacl like this:

------------[ cut here ]------------
refcount_t: underflow; use-after-free.
WARNING: CPU: 26 PID: 3199 at lib/refcount.c:28
refcount_warn_saturate+0xb5/0x170
Modules linked in:
CPU: 26 UID: 0 PID: 3199 Comm: nfsd Not tainted
6.12.0-rc6-00079-g04ae226af01f-dirty #8
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
1.16.1-2.fc37 04/01/2014
RIP: 0010:refcount_warn_saturate+0xb5/0x170
Code: cc cc 0f b6 1d b3 20 a5 03 80 fb 01 0f 87 65 48 d8 00 83 e3 01 75
e4 48 c7 c7 c0 3b 9b 85 c6 05 97 20 a5 03 01 e8 fb 3e 30 ff <0f> 0b eb
cd 0f b6 1d 8a3
RSP: 0018:ffffc90008637cd8 EFLAGS: 00010282
RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffffffff83904fde
RDX: dffffc0000000000 RSI: 0000000000000008 RDI: ffff88871ed36380
RBP: ffff888158beeb40 R08: 0000000000000001 R09: fffff520010c6f56
R10: ffffc90008637ab7 R11: 0000000000000001 R12: 0000000000000001
R13: ffff888140e77400 R14: ffff888140e77408 R15: ffffffff858b42c0
FS:  0000000000000000(0000) GS:ffff88871ed00000(0000)
knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000562384d32158 CR3: 000000055cc6a000 CR4: 00000000000006f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 
 ? refcount_warn_saturate+0xb5/0x170
 ? __warn+0xa5/0x140
 ? refcount_warn_saturate+0xb5/0x170
 ? report_bug+0x1b1/0x1e0
 ? handle_bug+0x53/0xa0
 ? exc_invalid_op+0x17/0x40
 ? asm_exc_invalid_op+0x1a/0x20
 ? tick_nohz_tick_stopped+0x1e/0x40
 ? refcount_warn_saturate+0xb5/0x170
 ? refcount_warn_saturate+0xb5/0x170
 nfs3svc_release_getacl+0xc9/0xe0
 svc_process_common+0x5db/0xb60
 ? __pfx_svc_process_common+0x10/0x10
 ? __rcu_read_unlock+0x69/0xa0
 ? __pfx_nfsd_dispatch+0x10/0x10
 ? svc_xprt_received+0xa1/0x120
 ? xdr_init_decode+0x11d/0x190
 svc_process+0x2a7/0x330
 svc_handle_xprt+0x69d/0x940
 svc_recv+0x180/0x2d0
 nfsd+0x168/0x200
 ? __pfx_nfsd+0x10/0x10
 kthread+0x1a2/0x1e0
 ? kthread+0xf4/0x1e0
 ? __pfx_kthread+0x10/0x10
 ret_from_fork+0x34/0x60
 ? __pfx_kthread+0x10/0x10
 ret_from_fork_asm+0x1a/0x30
 
Kernel panic - not syncing: kernel: panic_on_warn set ...

Clear acl_access/acl_default after posix_acl_release is called to prevent
UAF from being triggered.

Fixes: a257cdd0e217 ("[PATCH] NFSD: Add server support for NFSv3 ACLs.")
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/all/20241107014705.2509463-1-lilingfeng@huaweicloud.com/
Signed-off-by: Li Lingfeng 
Reviewed-by: Rick Macklem 
Reviewed-by: Jeff Layton 
Signed-off-by: Chuck Lever

Merge tag 'nfs-for-6.14-1' of git://git.linux-nfs.org/projects/anna/linux-nfs

2025-01-28T22:23:46+00:00

Pull NFS client updates from Anna Schumaker:
 "New Features:
   - Enable using direct IO with localio
   - Added localio related tracepoints

  Bugfixes:
   - Sunrpc fixes for working with a very large cl_tasks list
   - Fix a possible buffer overflow in nfs_sysfs_link_rpc_client()
   - Fixes for handling reconnections with localio
   - Fix how the NFS_FSCACHE kconfig option interacts with NETFS_SUPPORT
   - Fix COPY_NOTIFY xdr_buf size calculations
   - pNFS/Flexfiles fix for retrying requesting a layout segment for
     reads
   - Sunrpc fix for retrying on EKEYEXPIRED error when the TGT is
     expired

  Cleanups:
   - Various other nfs & nfsd localio cleanups
   - Prepratory patches for async copy improvements that are under
     development
   - Make OFFLOAD_CANCEL, LAYOUTSTATS, and LAYOUTERR moveable to other
     xprts
   - Add netns inum and srcaddr to debugfs rpc_xprt info"

* tag 'nfs-for-6.14-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (28 commits)
  SUNRPC: do not retry on EKEYEXPIRED when user TGT ticket expired
  sunrpc: add netns inum and srcaddr to debugfs rpc_xprt info
  pnfs/flexfiles: retry getting layout segment for reads
  NFSv4.2: make LAYOUTSTATS and LAYOUTERROR MOVEABLE
  NFSv4.2: mark OFFLOAD_CANCEL MOVEABLE
  NFSv4.2: fix COPY_NOTIFY xdr buf size calculation
  NFS: Rename struct nfs4_offloadcancel_data
  NFS: Fix typo in OFFLOAD_CANCEL comment
  NFS: CB_OFFLOAD can return NFS4ERR_DELAY
  nfs: Make NFS_FSCACHE select NETFS_SUPPORT instead of depending on it
  nfs: fix incorrect error handling in LOCALIO
  nfs: probe for LOCALIO when v3 client reconnects to server
  nfs: probe for LOCALIO when v4 client reconnects to server
  nfs/localio: remove redundant code and simplify LOCALIO enablement
  nfs_common: add nfs_localio trace events
  nfs_common: track all open nfsd_files per LOCALIO nfs_client
  nfs_common: rename nfslocalio nfs_uuid_lock to nfs_uuids_lock
  nfsd: nfsd_file_acquire_local no longer returns GC'd nfsd_file
  nfsd: rename nfsd_serv_ prefixed methods and variables with nfsd_net_
  nfsd: update percpu_ref to manage references on nfsd_net
  ...

Merge tag 'nfsd-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux

2025-01-28T01:27:24+00:00

Pull nfsd updates from Chuck Lever:
 "Jeff Layton contributed an implementation of NFSv4.2+ attribute
  delegation, as described here:

    https://www.ietf.org/archive/id/draft-ietf-nfsv4-delstid-08.html

  This interoperates with similar functionality introduced into the
  Linux NFS client in v6.11. An attribute delegation permits an NFS
  client to manage a file's mtime, rather than flushing dirty data to
  the NFS server so that the file's mtime reflects the last write, which
  is considerably slower.

  Neil Brown contributed dynamic NFSv4.1 session slot table resizing.
  This facility enables NFSD to increase or decrease the number of slots
  per NFS session depending on server memory availability. More session
  slots means greater parallelism.

  Chuck Lever fixed a long-standing latent bug where NFSv4 COMPOUND
  encoding screws up when crossing a page boundary in the encoding
  buffer. This is a zero-day bug, but hitting it is rare and depends on
  the NFS client implementation. The Linux NFS client does not happen to
  trigger this issue.

  A variety of bug fixes and other incremental improvements fill out the
  list of commits in this release. Great thanks to all contributors,
  reviewers, testers, and bug reporters who participated during this
  development cycle"

* tag 'nfsd-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (42 commits)
  sunrpc: Remove gss_{de,en}crypt_xdr_buf deadcode
  sunrpc: Remove gss_generic_token deadcode
  sunrpc: Remove unused xprt_iter_get_xprt
  Revert "SUNRPC: Reduce thread wake-up rate when receiving large RPC messages"
  nfsd: implement OPEN_ARGS_SHARE_ACCESS_WANT_OPEN_XOR_DELEGATION
  nfsd: handle delegated timestamps in SETATTR
  nfsd: add support for delegated timestamps
  nfsd: rework NFS4_SHARE_WANT_* flag handling
  nfsd: add support for FATTR4_OPEN_ARGUMENTS
  nfsd: prepare delegation code for handing out *_ATTRS_DELEG delegations
  nfsd: rename NFS4_SHARE_WANT_* constants to OPEN4_SHARE_ACCESS_WANT_*
  nfsd: switch to autogenerated definitions for open_delegation_type4
  nfs_common: make include/linux/nfs4.h include generated nfs4_1.h
  nfsd: fix handling of delegated change attr in CB_GETATTR
  SUNRPC: Document validity guarantees of the pointer returned by reserve_space
  NFSD: Insulate nfsd4_encode_fattr4() from page boundaries in the encode buffer
  NFSD: Insulate nfsd4_encode_secinfo() from page boundaries in the encode buffer
  NFSD: Refactor nfsd4_do_encode_secinfo() again
  NFSD: Insulate nfsd4_encode_readlink() from page boundaries in the encode buffer
  NFSD: Insulate nfsd4_encode_read_plus_data() from page boundaries in the encode buffer
  ...

Merge tag 'lsm-pr-20250121' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm

2025-01-22T04:03:04+00:00

Pull lsm updates from Paul Moore:

 - Improved handling of LSM "secctx" strings through lsm_context struct

   The LSM secctx string interface is from an older time when only one
   LSM was supported, migrate over to the lsm_context struct to better
   support the different LSMs we now have and make it easier to support
   new LSMs in the future.

   These changes explain the Rust, VFS, and networking changes in the
   diffstat.

 - Only build lsm_audit.c if CONFIG_SECURITY and CONFIG_AUDIT are
   enabled

   Small tweak to be a bit smarter about when we build the LSM's common
   audit helpers.

 - Check for absurdly large policies from userspace in SafeSetID

   SafeSetID policies rules are fairly small, basically just "UID:UID",
   it easy to impose a limit of KMALLOC_MAX_SIZE on policy writes which
   helps quiet a number of syzbot related issues. While work is being
   done to address the syzbot issues through other mechanisms, this is a
   trivial and relatively safe fix that we can do now.

 - Various minor improvements and cleanups

   A collection of improvements to the kernel selftests, constification
   of some function parameters, removing redundant assignments, and
   local variable renames to improve readability.

* tag 'lsm-pr-20250121' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm:
  lockdown: initialize local array before use to quiet static analysis
  safesetid: check size of policy writes
  net: corrections for security_secid_to_secctx returns
  lsm: rename variable to avoid shadowing
  lsm: constify function parameters
  security: remove redundant assignment to return variable
  lsm: Only build lsm_audit.c if CONFIG_SECURITY and CONFIG_AUDIT are set
  selftests: refactor the lsm `flags_overset_lsm_set_self_attr` test
  binder: initialize lsm_context structure
  rust: replace lsm context+len with lsm_context
  lsm: secctx provider check on release
  lsm: lsm_context in security_dentry_init_security
  lsm: use lsm_context in security_inode_getsecctx
  lsm: replace context+len with lsm_context
  lsm: ensure the correct LSM context releaser