linux.git/net/sunrpc, branch v6.6

SUNRPC/TLS: Lock the lower_xprt during the tls handshake

2023-09-27T19:16:40+00:00

Otherwise we run the risk of having the lower_xprt freed from underneath
us, causing an oops that looks like this:

[  224.150698] BUG: kernel NULL pointer dereference, address: 0000000000000018
[  224.150951] #PF: supervisor read access in kernel mode
[  224.151117] #PF: error_code(0x0000) - not-present page
[  224.151278] PGD 0 P4D 0
[  224.151361] Oops: 0000 [#1] PREEMPT SMP NOPTI
[  224.151499] CPU: 2 PID: 99 Comm: kworker/u10:6 Not tainted 6.6.0-rc3-g6465e260f487 #41264 a00b0960990fb7bc6d6a330ee03588b67f08a47b
[  224.151977] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS unknown 2/2/2022
[  224.152216] Workqueue: xprtiod xs_tcp_tls_setup_socket [sunrpc]
[  224.152434] RIP: 0010:xs_tcp_tls_setup_socket+0x3cc/0x7e0 [sunrpc]
[  224.152643] Code: 00 00 48 8b 7c 24 08 e9 f3 01 00 00 48 83 7b c0 00 0f 85 d2 01 00 00 49 8d 84 24 f8 05 00 00 48 89 44 24 10 48 8b 00 48 89 c5 <4c> 8b 68 18 66 41 83 3f 0a 75 71 45 31 ff 4c 89 ef 31 f6 e8 5c 76
[  224.153246] RSP: 0018:ffffb00ec060fd18 EFLAGS: 00010246
[  224.153427] RAX: 0000000000000000 RBX: ffff8c06c2e53e40 RCX: 0000000000000001
[  224.153652] RDX: ffff8c073bca2408 RSI: 0000000000000282 RDI: ffff8c06c259ee00
[  224.153868] RBP: 0000000000000000 R08: ffffffff9da55aa0 R09: 0000000000000001
[  224.154084] R10: 00000034306c30f1 R11: 0000000000000002 R12: ffff8c06c2e51800
[  224.154300] R13: ffff8c06c355d400 R14: 0000000004208160 R15: ffff8c06c2e53820
[  224.154521] FS:  0000000000000000(0000) GS:ffff8c073bd00000(0000) knlGS:0000000000000000
[  224.154763] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  224.154940] CR2: 0000000000000018 CR3: 0000000062c1e000 CR4: 0000000000750ee0
[  224.155157] PKRU: 55555554
[  224.155244] Call Trace:
[  224.155325]  
[  224.155395]  ? __die_body+0x68/0xb0
[  224.155507]  ? page_fault_oops+0x34c/0x3a0
[  224.155635]  ? _raw_spin_unlock_irqrestore+0xe/0x40
[  224.155793]  ? exc_page_fault+0x7a/0x1b0
[  224.155916]  ? asm_exc_page_fault+0x26/0x30
[  224.156047]  ? xs_tcp_tls_setup_socket+0x3cc/0x7e0 [sunrpc ae3a15912ae37fd51dafbdbc2dbd069117f8f5c8]
[  224.156367]  ? xs_tcp_tls_setup_socket+0x2fe/0x7e0 [sunrpc ae3a15912ae37fd51dafbdbc2dbd069117f8f5c8]
[  224.156697]  ? __pfx_xs_tls_handshake_done+0x10/0x10 [sunrpc ae3a15912ae37fd51dafbdbc2dbd069117f8f5c8]
[  224.157013]  process_scheduled_works+0x24e/0x450
[  224.157158]  worker_thread+0x21c/0x2d0
[  224.157275]  ? __pfx_worker_thread+0x10/0x10
[  224.157409]  kthread+0xe8/0x110
[  224.157510]  ? __pfx_kthread+0x10/0x10
[  224.157628]  ret_from_fork+0x37/0x50
[  224.157741]  ? __pfx_kthread+0x10/0x10
[  224.157859]  ret_from_fork_asm+0x1b/0x30
[  224.157983]  

Reviewed-by: Chuck Lever 
Signed-off-by: Anna Schumaker

Revert "SUNRPC dont update timeout value on connection reset"

2023-09-27T19:16:40+00:00

This reverts commit 88428cc4ae7abcc879295fbb19373dd76aad2bdd.

The problem this commit is intended to fix was comprehensively fixed
in commit 7de62bc09fe6 ("SUNRPC dont update timeout value on connection
reset").
Since then, this commit has been preventing the correct timeout of soft
mounted requests.

Cc: stable@vger.kernel.org # 5.9.x: 09252177d5f9: SUNRPC: Handle major timeout in xprt_adjust_timeout()
Cc: stable@vger.kernel.org # 5.9.x: 7de62bc09fe6: SUNRPC dont update timeout value on connection reset
Cc: stable@vger.kernel.org # 5.9.x
Signed-off-by: Trond Myklebust 
Signed-off-by: Anna Schumaker

SUNRPC: Fail quickly when server does not recognize TLS

2023-09-27T19:16:40+00:00

rpcauth_checkverf() should return a distinct error code when a
server recognizes the AUTH_TLS probe but does not support TLS so
that the client's header decoder can respond appropriately and
quickly. No retries are necessary is in this case, since the server
has already affirmatively answered "TLS is unsupported".

Suggested-by: Trond Myklebust 
Signed-off-by: Chuck Lever 
Signed-off-by: Anna Schumaker

NFSv4.1: fix pnfs MDS=DS session trunking

2023-09-13T15:51:11+00:00

Currently, when GETDEVICEINFO returns multiple locations where each
is a different IP but the server's identity is same as MDS, then
nfs4_set_ds_client() finds the existing nfs_client structure which
has the MDS's max_connect value (and if it's 1), then the 1st IP
on the DS's list will get dropped due to MDS trunking rules. Other
IPs would be added as they fall under the pnfs trunking rules.

For the list of IPs the 1st goes thru calling nfs4_set_ds_client()
which will eventually call nfs4_add_trunk() and call into
rpc_clnt_test_and_add_xprt() which has the check for MDS trunking.
The other IPs (after the 1st one), would call rpc_clnt_add_xprt()
which doesn't go thru that check.

nfs4_add_trunk() is called when MDS trunking is happening and it
needs to enforce the usage of max_connect mount option of the
1st mount. However, this shouldn't be applied to pnfs flow.

Instead, this patch proposed to treat MDS=DS as DS trunking and
make sure that MDS's max_connect limit does not apply to the
1st IP returned in the GETDEVICEINFO list. It does so by
marking the newly created client with a new flag NFS_CS_PNFS
which then used to pass max_connect value to use into the
rpc_clnt_test_and_add_xprt() instead of the existing rpc
client's max_connect value set by the MDS connection.

For example, mount was done without max_connect value set
so MDS's rpc client has cl_max_connect=1. Upon calling into
rpc_clnt_test_and_add_xprt() and using rpc client's value,
the caller passes in max_connect value which is previously
been set in the pnfs path (as a part of handling
GETDEVICEINFO list of IPs) in nfs4_set_ds_client().

However, when NFS_CS_PNFS flag is not set and we know we
are doing MDS trunking, comparing a new IP of the same
server, we then set the max_connect value to the
existing MDS's value and pass that into
rpc_clnt_test_and_add_xprt().

Fixes: dc48e0abee24 ("SUNRPC enforce creation of no more than max_connect xprts")
Signed-off-by: Olga Kornievskaia 
Signed-off-by: Anna Schumaker

Revert "SUNRPC: Fail faster on bad verifier"

2023-09-13T15:51:11+00:00

This reverts commit 0701214cd6e66585a999b132eb72ae0489beb724.

The premise of this commit was incorrect. There are exactly 2 cases
where rpcauth_checkverf() will return an error:

1) If there was an XDR decode problem (i.e. garbage data).
2) If gss_validate() had a problem verifying the RPCSEC_GSS MIC.

In the second case, there are again 2 subcases:

a) The GSS context expires, in which case gss_validate() will force a
   new context negotiation on retry by invalidating the cred.
b) The sequence number check failed because an RPC call timed out, and
   the client retransmitted the request using a new sequence number,
   as required by RFC2203.

In neither subcase is this a fatal error.

Reported-by: Russell Cattelan 
Fixes: 0701214cd6e6 ("SUNRPC: Fail faster on bad verifier")
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust 
Signed-off-by: Anna Schumaker

SUNRPC: Mark the cred for revalidation if the server rejects it

2023-09-13T15:51:11+00:00

If the server rejects the credential as being stale, or bad, then we
should mark it for revalidation before retransmitting.

Fixes: 7f5667a5f8c4 ("SUNRPC: Clean up rpc_verify_header()")
Signed-off-by: Trond Myklebust 
Signed-off-by: Anna Schumaker

Merge tag 'nfs-for-6.6-1' of git://git.linux-nfs.org/projects/anna/linux-nfs

2023-08-31T22:36:41+00:00

Pull NFS client updates from Anna Schumaker:
 "New Features:
   - Enable the NFS v4.2 READ_PLUS operation by default

  Stable Fixes:
   - NFSv4/pnfs: minor fix for cleanup path in nfs4_get_device_info
   - NFS: Fix a potential data corruption

  Bugfixes:
   - Fix various READ_PLUS issues including:
      - smatch warnings
      - xdr size calculations
      - scratch buffer handling
      - 32bit / highmem xdr page handling
   - Fix checkpatch errors in file.c
   - Fix redundant readdir request after an EOF
   - Fix handling of COPY ERR_OFFLOAD_NO_REQ
   - Fix assignment of xprtdata.cred

  Cleanups:
   - Remove unused xprtrdma function declarations
   - Clean up an integer overflow check to avoid a warning
   - Clean up #includes in dns_resolve.c
   - Clean up nfs4_get_device_info so we don't pass a NULL pointer
     to __free_page()
   - Clean up sunrpc TCP socket timeout configuration
   - Guard against READDIR loops when entry names are too long
   - Use EXCHID4_FLAG_USE_PNFS_DS for DS servers"

* tag 'nfs-for-6.6-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (22 commits)
  pNFS: Fix assignment of xprtdata.cred
  NFSv4.2: fix handling of COPY ERR_OFFLOAD_NO_REQ
  NFS: Guard against READDIR loop when entry names exceed MAXNAMELEN
  NFSv4.1: use EXCHGID4_FLAG_USE_PNFS_DS for DS server
  NFS/pNFS: Set the connect timeout for the pNFS flexfiles driver
  SUNRPC: Don't override connect timeouts in rpc_clnt_add_xprt()
  SUNRPC: Allow specification of TCP client connect timeout at setup
  SUNRPC: Refactor and simplify connect timeout
  SUNRPC: Set the TCP_SYNCNT to match the socket timeout
  NFS: Fix a potential data corruption
  nfs: fix redundant readdir request after get eof
  nfs/blocklayout: Use the passed in gfp flags
  filemap: Fix errors in file.c
  NFSv4/pnfs: minor fix for cleanup path in nfs4_get_device_info
  NFS: Move common includes outside ifdef
  SUNRPC: clean up integer overflow check
  xprtrdma: Remove unused function declaration rpcrdma_bc_post_recv()
  NFS: Enable the READ_PLUS operation by default
  SUNRPC: kmap() the xdr pages during decode
  NFSv4.2: Rework scratch handling for READ_PLUS (again)
  ...

Merge tag 'nfsd-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux

2023-08-31T22:32:18+00:00

Pull nfsd updates from Chuck Lever:
 "I'm thrilled to announce that the Linux in-kernel NFS server now
  offers NFSv4 write delegations. A write delegation enables a client to
  cache data and metadata for a single file more aggressively, reducing
  network round trips and server workload. Many thanks to Dai Ngo for
  contributing this facility, and to Jeff Layton and Neil Brown for
  reviewing and testing it.

  This release also sees the removal of all support for DES- and
  triple-DES-based Kerberos encryption types in the kernel's SunRPC
  implementation. These encryption types have been deprecated by the
  Internet community for years and are considered insecure. This change
  affects both the in-kernel NFS client and server.

  The server's UDP and TCP socket transports have now fully adopted
  David Howells' new bio_vec iterator so that no more than one sendmsg()
  call is needed to transmit each RPC message. In particular, this helps
  kTLS optimize record boundaries when sending RPC-with-TLS replies, and
  it takes the server a baby step closer to handling file I/O via
  folios.

  We've begun work on overhauling the SunRPC thread scheduler to remove
  a costly linked-list walk when looking for an idle RPC service thread
  to wake. The pre-requisites are included in this release. Thanks to
  Neil Brown for his ongoing work on this improvement"

* tag 'nfsd-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux: (56 commits)
  Documentation: Add missing documentation for EXPORT_OP flags
  SUNRPC: Remove unused declaration rpc_modcount()
  SUNRPC: Remove unused declarations
  NFSD: da_addr_body field missing in some GETDEVICEINFO replies
  SUNRPC: Remove return value of svc_pool_wake_idle_thread()
  SUNRPC: make rqst_should_sleep() idempotent()
  SUNRPC: Clean up svc_set_num_threads
  SUNRPC: Count ingress RPC messages per svc_pool
  SUNRPC: Deduplicate thread wake-up code
  SUNRPC: Move trace_svc_xprt_enqueue
  SUNRPC: Add enum svc_auth_status
  SUNRPC: change svc_xprt::xpt_flags bits to enum
  SUNRPC: change svc_rqst::rq_flags bits to enum
  SUNRPC: change svc_pool::sp_flags bits to enum
  SUNRPC: change cache_head.flags bits to enum
  SUNRPC: remove timeout arg from svc_recv()
  SUNRPC: change svc_recv() to return void.
  SUNRPC: call svc_process() from svc_recv().
  nfsd: separate nfsd_last_thread() from nfsd_put()
  nfsd: Simplify code around svc_exit_thread() call in nfsd()
  ...

SUNRPC: Remove return value of svc_pool_wake_idle_thread()

2023-08-29T21:45:22+00:00

The returned value is not used (any more), so don't return it.

Signed-off-by: NeilBrown 
Signed-off-by: Chuck Lever

SUNRPC: make rqst_should_sleep() idempotent()

2023-08-29T21:45:22+00:00

Based on its name you would think that rqst_should_sleep() would be
read-only, not changing anything.  But in fact it will clear
SP_TASK_PENDING if that was set.  This is surprising, and it blurs the
line between "check for work to do" and "dequeue work to do".

So change the "test_and_clear" to simple "test" and clear the bit once
the thread has decided to wake up and return to the caller.

With this, it makes sense to *always* set SP_TASK_PENDING when asked,
rather than to set it only if no thread could be woken up.

[ cel: Previously TASK_PENDING indicated there is work waiting but no
idle threads were found to pick up that work. After this patch, it acts
as an XPT_BUSY flag for wake-ups that have no associated xprt. ]

Signed-off-by: NeilBrown 
Reviewed-by: Jeff Layton 
Signed-off-by: Chuck Lever