Commit Graph

23 Commits

Author SHA1 Message Date
Trond Myklebust fc821d5920 pnfs/NFSv4.1: Add multipath capabilities to pNFS flexfiles servers over NFSv3
This adds multipathing to pNFS over NFSv3 as described in the flexfiles
draft spec.
Ideally, we'd like to do the same for pNFS files, but the NFSv4.1 protocol
requires a call to EXCHANGE_ID in order to test that the connection can do
session trunking.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2016-02-05 18:48:57 -05:00
Trond Myklebust b20135d0b2 NFSv4.1/pNFS: Don't queue up a new commit if the layout segment is invalid
If the layout segment is invalid, then we should not be adding more
write requests to the commit list. Instead, those writes should be
replayed after requesting a new layout.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-31 15:55:35 -05:00
Trond Myklebust af7cf05793 NFS: Allow multiple commit requests in flight per file
Allow synchronous RPC calls to wait for pending RPC calls to finish,
but also allow asynchronous ones to just fire off another commit.

With this patch, the xfstests generic/074 test completes in 226s
instead of 242s

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-31 13:53:48 -05:00
Peng Tao 046be74da8 NFS41: fix list splice type
We want to move commiting pages to pages list instead.
Otherwise it causes pnfs small writes crash like:

[34560.037692] BUG: unable to handle kernel NULL pointer dereference at 0000000000000068
[34560.038557] IP: [<ffffffffa05423d6>] nfs_init_commit+0x26/0x130 [nfs]
[34560.039400] PGD 69f5a067 PUD 69f59067 PMD 0
[34560.040207] Oops: 0000 [#1] SMP
[34560.041014] Modules linked in: nfsv3(OE) nfs_layout_flexfiles(OE) nfsv4(OE) nfs(OE) fscache(E) rpcsec_gss_krb5(E) xt_addrtype(E) xt_conntrack(E) ipt_MASQUERADE(E) nf_nat_masquerade_ipv4(E) iptable_nat(E) nf_conntrack_ipv4(E) nf_defrag_ipv4(E) nf_nat_ipv4(E) iptable_filter(E) ip_tables(E) x_tables(E) nf_nat(E) nf_conntrack(E) bridge(E) stp(E) llc(E) dm_thin_pool(E) dm_persistent_data(E) dm_bio_prison(E) dm_bufio(E) ppdev(E) vmw_balloon(E) coretemp(E) crc32_pclmul(E) ghash_clmulni_intel(E) aesni_intel(E) aes_x86_64(E) glue_helper(E) lrw(E) gf128mul(E) ablk_helper(E) cryptd(E) psmouse(E) serio_raw(E) vmw_vmci(E) i2c_piix4(E) shpchp(E) parport_pc(E) parport(E) mac_hid(E) nfsd(E) auth_rpcgss(E) nfs_acl(E) lockd(E) grace(E) sunrpc(E) xfs(E) libcrc32c(E) hid_generic(E) usbhid(E) hid(E) e1000(E) mptspi(E)
[34560.045106]  mptscsih(E) mptbase(E) vmwgfx(E) drm_kms_helper(E) ttm(E) drm(E) autofs4(E) [last unloaded: fscache]
[34560.045897] CPU: 0 PID: 130543 Comm: bash Tainted: G           OE   4.2.0-rc5-dp-00057-gf993a93 #11
[34560.046699] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 05/20/2014
[34560.047525] task: ffff880031b0a980 ti: ffff880045fec000 task.ti: ffff880045fec000
[34560.048264] RIP: 0010:[<ffffffffa05423d6>]  [<ffffffffa05423d6>] nfs_init_commit+0x26/0x130 [nfs]
[34560.049000] RSP: 0018:ffff880045fefc18  EFLAGS: 00010246
[34560.049717] RAX: 0000000000000000 RBX: ffff8800208fbc80 RCX: ffff880045fefd50
[34560.050396] RDX: ffff880031c19ec0 RSI: ffff880045fefc88 RDI: ffff8800208fbc80
[34560.051041] RBP: ffff880045fefc28 R08: ffff8800208fbe68 R09: ffff880045fefc88
[34560.051666] R10: 0000000000000000 R11: 0000000000000000 R12: ffff880045fefc78
[34560.052247] R13: ffff880045fefc88 R14: ffff880045fefa90 R15: ffff880045fefd50
[34560.052825] FS:  00007fa02d58c740(0000) GS:ffff88006d600000(0000) knlGS:0000000000000000
[34560.053410] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[34560.053992] CR2: 0000000000000068 CR3: 000000003b37a000 CR4: 00000000001406f0
[34560.054615] Stack:
[34560.055200]  ffff8800208fbc80 ffff8800208fbc80 ffff880045fefcc8 ffffffffa05c1a5b
[34560.055800]  ffff880045fefcc8 ffff880045fefd50 0000000045fefcb8 ffff880045fefd40
[34560.056418]  ffff8800420608e0 ffffffffa04f3910 0000000100000001 ffff880045fefd50
[34560.057013] Call Trace:
[34560.057672]  [<ffffffffa05c1a5b>] pnfs_generic_commit_pagelist+0x1cb/0x300 [nfsv4]
[34560.058277]  [<ffffffffa04f3910>] ? ff_layout_commit_pagelist+0x20/0x20 [nfs_layout_flexfiles]
[34560.058907]  [<ffffffffa04f3905>] ff_layout_commit_pagelist+0x15/0x20 [nfs_layout_flexfiles]
[34560.059557]  [<ffffffffa0543fc1>] nfs_generic_commit_list+0xb1/0xf0 [nfs]
[34560.060214]  [<ffffffffa0543e47>] ? nfs_scan_commit+0x37/0xa0 [nfs]
[34560.060825]  [<ffffffffa0544081>] nfs_commit_inode+0x81/0x150 [nfs]
[34560.061432]  [<ffffffffa05443ae>] nfs_wb_all+0x1ae/0x400 [nfs]
[34560.062035]  [<ffffffffa05380ad>] nfs_getattr+0x33d/0x510 [nfs]
[34560.062630]  [<ffffffff8122499c>] vfs_getattr_nosec+0x2c/0x40
[34560.063223]  [<ffffffff81224a66>] vfs_getattr+0x26/0x30
[34560.063818]  [<ffffffff81224b35>] vfs_fstatat+0x65/0xa0
[34560.064413]  [<ffffffff81224f3f>] SYSC_newstat+0x1f/0x40
[34560.065016]  [<ffffffff8102b176>] ? do_audit_syscall_entry+0x66/0x70
[34560.065626]  [<ffffffff8102c773>] ? syscall_trace_enter_phase1+0x113/0x170
[34560.066245]  [<ffffffff81003017>] ? trace_hardirqs_on_thunk+0x17/0x19
[34560.066868]  [<ffffffff812251ae>] SyS_newstat+0xe/0x10
[34560.067533]  [<ffffffff817a5df2>] entry_SYSCALL_64_fastpath+0x16/0x7a
[34560.068173] Code: 0f 1f 44 00 00 0f 1f 44 00 00 55 4c 8d 87 e8 01 00 00 48 89 e5 53 48 89 fb 48 83 ec 08 4c 8b 0e 49 8b 41 18 4c 39 ce 48 8b 40 40 <4c> 8b 50 68 74 24 48 8b 87 e8 01 00 00 48 8b 7e 08 4d 89 41 08
[34560.069609] RIP  [<ffffffffa05423d6>] nfs_init_commit+0x26/0x130 [nfs]
[34560.070295]  RSP <ffff880045fefc18>
[34560.071008] CR2: 0000000000000068
[34560.073207] ---[ end trace f85f873260977406 ]---

[fixes 27571297a7e(pNFS: Tighten up locking around DS commit buckets)]
Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-20 13:43:53 -05:00
Trond Myklebust 74918f937d Merge branch 'layoutfixes'
* layoutfixes:
  NFSv4.1/pnfs: Remove redundant wakeup in pnfs_send_layoutreturn()
  NFSv4.1/pnfs: Remove redundant check in pnfs_layoutgets_blocked()
  NFSv4.1/pnfs: Remove redundant lo->plh_block_lgets in layoutreturn
  NFSv4.1/pnfs: Don't prevent layoutgets when doing return-on-close
  NFSv4.1/pnfs: Fix serialisation of layout return and layoutget
  NFSv4.1/pnfs: Remove redundant checks in pnfs_layoutgets_blocked()
  pNFS: Tighten up locking around DS commit buckets
2015-08-17 13:36:32 -05:00
Trond Myklebust 6f536936b7 NFSv4.1/pNFS: Fix borken function _same_data_server_addrs_locked()
- Switch back to using list_for_each_entry(). Fixes an incorrect test
  for list NULL termination.
- Do not assume that lists are sorted.
- Finally, consider an existing entry to match if it consists of a subset
  of the addresses in the new entry.

Cc: stable@vger.kernel.org # 4.0+
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-17 13:05:43 -05:00
Trond Myklebust 27571297a7 pNFS: Tighten up locking around DS commit buckets
I'm not aware of any bugreports around this issue, but the locking
around the pnfs_commit_bucket is inconsistent at best. This patch
tightens it up by ensuring that the 'bucket->committing' list is always
changed atomically w.r.t. the 'bucket->clseg' layout segment tracking.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-12 14:56:18 -04:00
Trond Myklebust 86d80f9734 NFSv4.1/pnfs: Fix atomicity of commit list updates
pnfs_layout_mark_request_commit() needs to ensure that it adds the
request to the commit list atomically with all the other updates
in order to prevent corruption to buckets[ds_commit_idx].wlseg
due to races with pnfs_generic_clear_request_commit().

Fixes: 338d00cfef ("pnfs: Refactor the *_layout_mark_request_commit...")
Cc: stable@vger.kernel.org # v4.0+
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-10 19:08:13 -04:00
Trond Myklebust 21330b6670 Merge branch 'bugfixes'
* bugfixes:
  NFSv4: Return delegations synchronously in evict_inode
  SUNRPC: Fix a regression when reconnecting
  NFS: remount with security change should return EINVAL
  nfs: do not export discarded symbols
  NFSv4.1: don't export static symbol
2015-04-23 15:16:27 -04:00
Trond Myklebust 5bb89b4702 NFSv4.1/pnfs: Separate out metadata and data consistency for pNFS
The LAYOUTCOMMIT operation means different things to different layout types.
For blocks and objects, it is both a data and metadata consistency operation.
For files and flexfiles, it is only a metadata consistency operation.

This patch separates out the 2 cases, allowing the files/flexfiles layout
drivers to optimise away the data consistency calls to layoutcommit.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-03-27 12:39:38 -04:00
Arnd Bergmann df137bc125 nfs: do not export discarded symbols
The function nfs4_pnfs_v3_ds_connect_unload is exported so it can be
used by other modules, but is also marked '__exit' and will be
discarded when built into the kernel, as pointed out by this
linker error:

`nfs4_pnfs_v3_ds_connect_unload' referenced in section `___ksymtab_gpl+nfs4_pnfs_v3_ds_connect_unload' of fs/built-in.o: defined in discarded section `.exit.text' of fs/built-in.o

This removes the __exit annotation to make it safe to call this function.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 5f01d95394 ("nfs41: create NFSv3 DS connection if specified")
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-03-12 11:53:24 -04:00
Tom Haynes 338d00cfef pnfs: Refactor the *_layout_mark_request_commit to use pnfs_layout_mark_request_commit
The File Layout's filelayout_mark_request_commit() is almost the
Flex File Layout's ff_layout_mark_request_commit(). And that can
be reduced by calling into nfs_request_add_commit_list().

Signed-off-by: Tom Haynes <loghyr@primarydata.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-02-18 07:20:35 -08:00
Weston Andros Adamson b57ff1303a pnfs: pass ds_commit_idx through the commit path
Pass ds_commit_idx through the nfs commit path. It's used to select
the commit bucket when using pnfs and is ignored when not using pnfs.
Several functions had to be changed: nfs_retry_commit,
nfs_mark_request_commit, pnfs_mark_request_commit and the pnfs layout
driver .mark_request_commit functions.

Signed-off-by: Tom Haynes <loghyr@primarydata.com>
2015-02-03 11:06:45 -08:00
Peng Tao c220106fb4 nfs/filelayout: use pnfs_error_mark_layout_for_return
Instead of calling layoutreturn directly, call pnfs_error_mark_layout_for_return
to mark layouts for return and let generic code return layout when
layout segments are freed.

Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>

Conflicts:
	fs/nfs/filelayout/filelayout.c
2015-02-03 11:06:43 -08:00
Peng Tao c36aae9ad9 nfs: allow different protocol in nfs_initiate_commit
pnfs flexfile layout client may want to use NFSv3 ops rather
than the default MDS v4 ops.

Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:35 -08:00
Peng Tao 5f01d95394 nfs41: create NFSv3 DS connection if specified
Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:35 -08:00
Peng Tao 30626f9c32 nfs41: allow LD to choose DS connection version/minor_version
flexfile layout may need to set such when making DS connections.

Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:34 -08:00
Peng Tao 064172f345 nfs41: allow LD to choose DS connection auth flavor
flexfile layout may use different auth flavor as specified by MDS.

Reviewed-by: Jeff Layton <jlayton@primarydata.com>
Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:33 -08:00
Peng Tao 7405f9e195 nfs41: pull nfs4_ds_connect from file layout to generic pnfs
It can be reused by flexfiles layout client.

Reviewed-by: Jeff Layton <jlayton@primarydata.com>
Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:32 -08:00
Peng Tao 6b7f3cf963 nfs41: pull decode_ds_addr from file layout to generic pnfs
It can be reused by flexfile layout.

Reviewed-by: Jeff Layton <jlayton@primarydata.com>
Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:32 -08:00
Peng Tao 875ae0694b nfs41: pull data server cache from file layout to generic pnfs
Also pull nfs4_pnfs_ds_addr and nfs4_pnfs_ds to generic pnfs.

They can all be reused by flexfile layout as well.

Reviewed-by: Jeff Layton <jlayton@primarydata.com>
Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Tom Haynes <Thomas.Haynes@primarydata.com>
2015-02-03 11:06:32 -08:00
Tom Haynes 085d1e33a6 pnfs: Do not grab the commit_info lock twice when rescheduling writes
Acked-by: Jeff Layton <jlayton@primarydata.com>
Signed-off-by: Tom Haynes <loghyr@primarydata.com>
2015-02-03 11:06:31 -08:00
Tom Haynes f54bcf2ece pnfs: Prepare for flexfiles by pulling out common code
The flexfilelayout driver will share some common code
with the filelayout driver. This set of changes refactors
that common code out to avoid any module depenencies.

Signed-off-by: Tom Haynes <loghyr@primarydata.com>
2015-02-03 11:06:31 -08:00