From 42d42a5b0cd263757f8e519debbc744fdaefdaf7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 23 May 2016 09:24:55 -0400 Subject: SUNRPC: Small optimisation of client receive Do not queue the client receive work if we're still processing. Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprtsock.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/sunrpc/xprtsock.h b/include/linux/sunrpc/xprtsock.h index 0ece4ba06f06..bef3fb0abb8f 100644 --- a/include/linux/sunrpc/xprtsock.h +++ b/include/linux/sunrpc/xprtsock.h @@ -80,6 +80,7 @@ struct sock_xprt { #define TCP_RPC_REPLY (1UL << 6) #define XPRT_SOCK_CONNECTING 1U +#define XPRT_SOCK_DATA_READY (2) #endif /* __KERNEL__ */ -- cgit v1.2.3 From 40a5f1b19bacb2de7a051be952dee85e38c9e5f5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 27 May 2016 10:39:50 -0400 Subject: SUNRPC: RPC transport queue must be low latency rpciod can easily get congested due to the long list of queued rpc_tasks. Having the receive queue wait in turn for those tasks to complete can therefore be a bottleneck. Address the problem by separating the workqueues into: - rpciod: manages rpc_tasks - xprtiod: manages transport related work. Signed-off-by: Trond Myklebust --- include/linux/sunrpc/sched.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h index 05a1809c44d9..ef780b3b5e31 100644 --- a/include/linux/sunrpc/sched.h +++ b/include/linux/sunrpc/sched.h @@ -247,6 +247,7 @@ void rpc_show_tasks(struct net *); int rpc_init_mempool(void); void rpc_destroy_mempool(void); extern struct workqueue_struct *rpciod_workqueue; +extern struct workqueue_struct *xprtiod_workqueue; void rpc_prepare_task(struct rpc_task *task); static inline int rpc_wait_for_completion_task(struct rpc_task *task) -- cgit v1.2.3 From f1dc237c60a5fdecc83062a28a702193f881cb19 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 27 May 2016 12:59:33 -0400 Subject: SUNRPC: Reduce latency when send queue is congested Use the low latency transport workqueue to process the task that is next in line on the xprt->sending queue. Signed-off-by: Trond Myklebust --- include/linux/sunrpc/sched.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'include') diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h index ef780b3b5e31..817af0b4385e 100644 --- a/include/linux/sunrpc/sched.h +++ b/include/linux/sunrpc/sched.h @@ -230,6 +230,10 @@ void rpc_wake_up_queued_task(struct rpc_wait_queue *, struct rpc_task *); void rpc_wake_up(struct rpc_wait_queue *); struct rpc_task *rpc_wake_up_next(struct rpc_wait_queue *); +struct rpc_task *rpc_wake_up_first_on_wq(struct workqueue_struct *wq, + struct rpc_wait_queue *, + bool (*)(struct rpc_task *, void *), + void *); struct rpc_task *rpc_wake_up_first(struct rpc_wait_queue *, bool (*)(struct rpc_task *, void *), void *); -- cgit v1.2.3 From 6b56a89833fa7903595c8d138bb4927187315cba Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 1 Jun 2016 18:23:01 -0400 Subject: NFS: Kill NFS_INO_NFS_INO_FLUSHING: it is a performance killer filemap_datawrite() and friends already deal just fine with livelock. Signed-off-by: Trond Myklebust --- include/linux/nfs_fs.h | 1 - 1 file changed, 1 deletion(-) (limited to 'include') diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index d71278c3c5bd..120dd04b553c 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -205,7 +205,6 @@ struct nfs_inode { #define NFS_INO_STALE (1) /* possible stale inode */ #define NFS_INO_ACL_LRU_SET (2) /* Inode is on the LRU list */ #define NFS_INO_INVALIDATING (3) /* inode is being invalidated */ -#define NFS_INO_FLUSHING (4) /* inode is flushing out data */ #define NFS_INO_FSCACHE (5) /* inode can be cached by FS-Cache */ #define NFS_INO_FSCACHE_LOCK (6) /* FS-Cache cookie management lock */ #define NFS_INO_LAYOUTCOMMIT (9) /* layoutcommit required */ -- cgit v1.2.3 From 5c6e5b60aae4347223f176966455010a5715b863 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 22 Jun 2016 14:13:12 -0400 Subject: NFS: Fix an Oops in the pNFS files and flexfiles connection setup to the DS Chris Worley reports: RIP: 0010:[] [] rpc_new_client+0x2a0/0x2e0 [sunrpc] RSP: 0018:ffff880158f6f548 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff880234f8bc00 RCX: 000000000000ea60 RDX: 0000000000074cc0 RSI: 000000000000ea60 RDI: ffff880234f8bcf0 RBP: ffff880158f6f588 R08: 000000000001ac80 R09: ffff880237003300 R10: ffff880201171000 R11: ffffea0000d75200 R12: ffffffffa03afc60 R13: ffff880230c18800 R14: 0000000000000000 R15: ffff880158f6f680 FS: 00007f0e32673740(0000) GS:ffff88023fc40000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000008 CR3: 0000000234886000 CR4: 00000000001406e0 Stack: ffffffffa047a680 0000000000000000 ffff880158f6f598 ffff880158f6f680 ffff880158f6f680 ffff880234d11d00 ffff88023357f800 ffff880158f6f7d0 ffff880158f6f5b8 ffffffffa024660a ffff880158f6f5b8 ffffffffa02492ec Call Trace: [] rpc_create_xprt+0x1a/0xb0 [sunrpc] [] ? xprt_create_transport+0x13c/0x240 [sunrpc] [] rpc_create+0xc6/0x1a0 [sunrpc] [] nfs_create_rpc_client+0xf5/0x140 [nfs] [] nfs_init_client+0x3a/0xd0 [nfs] [] nfs_get_client+0x25f/0x310 [nfs] [] ? rpc_ntop+0xe8/0x100 [sunrpc] [] nfs3_set_ds_client+0xcc/0x100 [nfsv3] [] nfs4_pnfs_ds_connect+0x120/0x400 [nfsv4] [] nfs4_ff_layout_prepare_ds+0xe7/0x330 [nfs_layout_flexfiles] [] ff_layout_pg_init_write+0xcb/0x280 [nfs_layout_flexfiles] [] __nfs_pageio_add_request+0x12c/0x490 [nfs] [] nfs_pageio_add_request+0xc2/0x2a0 [nfs] [] ? nfs_pageio_init+0x75/0x120 [nfs] [] nfs_do_writepage+0x120/0x270 [nfs] [] nfs_writepage_locked+0x61/0xc0 [nfs] [] ? __percpu_counter_add+0x55/0x70 [] nfs_wb_single_page+0xef/0x1c0 [nfs] [] ? __dec_zone_page_state+0x33/0x40 [] nfs_launder_page+0x41/0x90 [nfs] [] invalidate_inode_pages2_range+0x340/0x3a0 [] invalidate_inode_pages2+0x17/0x20 [] nfs_release+0x9e/0xb0 [nfs] [] ? nfs_open+0x60/0x60 [nfs] [] nfs_file_release+0x3d/0x60 [nfs] [] __fput+0xdc/0x1e0 [] ____fput+0xe/0x10 [] task_work_run+0xc4/0xe0 [] do_exit+0x2e8/0xb30 [] ? do_audit_syscall_entry+0x6c/0x70 [] ? __audit_syscall_exit+0x1e6/0x280 [] do_group_exit+0x3f/0xa0 [] SyS_exit_group+0x14/0x20 [] system_call_fastpath+0x12/0x71 Which seems to be due to a call to utsname() when in a task exit context in order to determine the hostname to set in rpc_new_client(). In reality, what we want here is not the hostname of the current task, but the hostname that was used to set up the metadata server. Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index c304a11b5b1a..82b81a1c2438 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1596,9 +1596,8 @@ struct nfs_rpc_ops { int (*have_delegation)(struct inode *, fmode_t); int (*return_delegation)(struct inode *); struct nfs_client *(*alloc_client) (const struct nfs_client_initdata *); - struct nfs_client * - (*init_client) (struct nfs_client *, const struct rpc_timeout *, - const char *); + struct nfs_client *(*init_client) (struct nfs_client *, + const struct nfs_client_initdata *); void (*free_client) (struct nfs_client *); struct nfs_server *(*create_server)(struct nfs_mount_info *, struct nfs_subversion *); struct nfs_server *(*clone_server)(struct nfs_server *, struct nfs_fh *, -- cgit v1.2.3 From a5864c999de6703f7ce908f72337568520c6cad3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 3 Jun 2016 17:07:19 -0400 Subject: NFS: Do not serialise O_DIRECT reads and writes Allow dio requests to be scheduled in parallel, but ensuring that they do not conflict with buffered I/O. Signed-off-by: Trond Myklebust --- include/linux/nfs_fs.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 120dd04b553c..225d17d35277 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -210,6 +210,7 @@ struct nfs_inode { #define NFS_INO_LAYOUTCOMMIT (9) /* layoutcommit required */ #define NFS_INO_LAYOUTCOMMITTING (10) /* layoutcommit inflight */ #define NFS_INO_LAYOUTSTATS (11) /* layoutstats inflight */ +#define NFS_INO_ODIRECT (12) /* I/O setting is O_DIRECT */ static inline struct nfs_inode *NFS_I(const struct inode *inode) { -- cgit v1.2.3 From be527494e02b89e03485955b30de6c1e976a07eb Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 22 Jun 2016 08:19:36 -0400 Subject: NFS: Remove unused function nfs_revalidate_mapping_protected() Clean up... Signed-off-by: Trond Myklebust --- include/linux/nfs_fs.h | 1 - 1 file changed, 1 deletion(-) (limited to 'include') diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 225d17d35277..810124b33327 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -351,7 +351,6 @@ extern int nfs_revalidate_inode_rcu(struct nfs_server *server, struct inode *ino extern int __nfs_revalidate_inode(struct nfs_server *, struct inode *); extern int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping); extern int nfs_revalidate_mapping_rcu(struct inode *inode); -extern int nfs_revalidate_mapping_protected(struct inode *inode, struct address_space *mapping); extern int nfs_setattr(struct dentry *, struct iattr *); extern void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr, struct nfs_fattr *); extern void nfs_setsecurity(struct inode *inode, struct nfs_fattr *fattr, -- cgit v1.2.3 From 65b80179f9b8171b74625febf3457f41e792fa23 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 29 Jun 2016 13:55:06 -0400 Subject: xprtrdma: No direct data placement with krb5i and krb5p Direct data placement is not allowed when using flavors that guarantee integrity or privacy. When such security flavors are in effect, don't allow the use of Read and Write chunks for moving individual data items. All messages larger than the inline threshold are sent via Long Call or Long Reply. On my systems (CX-3 Pro on FDR), for small I/O operations, the use of Long messages adds only around 5 usecs of latency in each direction. Note that when integrity or encryption is used, the host CPU touches every byte in these messages. Even if it could be used, data movement offload doesn't buy much in this case. Signed-off-by: Chuck Lever Tested-by: Steve Wise Signed-off-by: Anna Schumaker --- include/linux/sunrpc/auth.h | 3 +++ include/linux/sunrpc/gss_api.h | 2 ++ 2 files changed, 5 insertions(+) (limited to 'include') diff --git a/include/linux/sunrpc/auth.h b/include/linux/sunrpc/auth.h index 899791573a40..3a40287b4d27 100644 --- a/include/linux/sunrpc/auth.h +++ b/include/linux/sunrpc/auth.h @@ -107,6 +107,9 @@ struct rpc_auth { /* per-flavor data */ }; +/* rpc_auth au_flags */ +#define RPCAUTH_AUTH_DATATOUCH 0x00000002 + struct rpc_auth_create_args { rpc_authflavor_t pseudoflavor; const char *target_name; diff --git a/include/linux/sunrpc/gss_api.h b/include/linux/sunrpc/gss_api.h index 1f911ccb2a75..68ec78c1aa48 100644 --- a/include/linux/sunrpc/gss_api.h +++ b/include/linux/sunrpc/gss_api.h @@ -73,6 +73,7 @@ u32 gss_delete_sec_context( rpc_authflavor_t gss_svc_to_pseudoflavor(struct gss_api_mech *, u32 qop, u32 service); u32 gss_pseudoflavor_to_service(struct gss_api_mech *, u32 pseudoflavor); +bool gss_pseudoflavor_to_datatouch(struct gss_api_mech *, u32 pseudoflavor); char *gss_service_to_auth_domain_name(struct gss_api_mech *, u32 service); struct pf_desc { @@ -81,6 +82,7 @@ struct pf_desc { u32 service; char *name; char *auth_domain_name; + bool datatouch; }; /* Different mechanisms (e.g., krb5 or spkm3) may implement gss-api, and -- cgit v1.2.3 From ce52914eb76efd62aa48d738cf845b37852bf920 Mon Sep 17 00:00:00 2001 From: Scott Mayhew Date: Tue, 7 Jun 2016 15:14:48 -0400 Subject: sunrpc: move NO_CRKEY_TIMEOUT to the auth->au_flags A generic_cred can be used to look up a unx_cred or a gss_cred, so it's not really safe to use the the generic_cred->acred->ac_flags to store the NO_CRKEY_TIMEOUT flag. A lookup for a unx_cred triggered while the KEY_EXPIRE_SOON flag is already set will cause both NO_CRKEY_TIMEOUT and KEY_EXPIRE_SOON to be set in the ac_flags, leaving the user associated with the auth_cred to be in a state where they're perpetually doing 4K NFS_FILE_SYNC writes. This can be reproduced as follows: 1. Mount two NFS filesystems, one with sec=krb5 and one with sec=sys. They do not need to be the same export, nor do they even need to be from the same NFS server. Also, v3 is fine. $ sudo mount -o v3,sec=krb5 server1:/export /mnt/krb5 $ sudo mount -o v3,sec=sys server2:/export /mnt/sys 2. As the normal user, before accessing the kerberized mount, kinit with a short lifetime (but not so short that renewing the ticket would leave you within the 4-minute window again by the time the original ticket expires), e.g. $ kinit -l 10m -r 60m 3. Do some I/O to the kerberized mount and verify that the writes are wsize, UNSTABLE: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 4. Wait until you're within 4 minutes of key expiry, then do some more I/O to the kerberized mount to ensure that RPC_CRED_KEY_EXPIRE_SOON gets set. Verify that the writes are 4K, FILE_SYNC: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 5. Now do some I/O to the sec=sys mount. This will cause RPC_CRED_NO_CRKEY_TIMEOUT to be set: $ dd if=/dev/zero of=/mnt/sys/file bs=1M count=1 6. Writes for that user will now be permanently 4K, FILE_SYNC for that user, regardless of which mount is being written to, until you reboot the client. Renewing the kerberos ticket (assuming it hasn't already expired) will have no effect. Grabbing a new kerberos ticket at this point will have no effect either. Move the flag to the auth->au_flags field (which is currently unused) and rename it slightly to reflect that it's no longer associated with the auth_cred->ac_flags. Add the rpc_auth to the arg list of rpcauth_cred_key_to_expire and check the au_flags there too. Finally, add the inode to the arg list of nfs_ctx_key_to_expire so we can determine the rpc_auth to pass to rpcauth_cred_key_to_expire. Signed-off-by: Scott Mayhew Signed-off-by: Trond Myklebust --- include/linux/sunrpc/auth.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/sunrpc/auth.h b/include/linux/sunrpc/auth.h index 899791573a40..f890a295a7ff 100644 --- a/include/linux/sunrpc/auth.h +++ b/include/linux/sunrpc/auth.h @@ -37,7 +37,6 @@ struct rpcsec_gss_info; /* auth_cred ac_flags bits */ enum { - RPC_CRED_NO_CRKEY_TIMEOUT = 0, /* underlying cred has no key timeout */ RPC_CRED_KEY_EXPIRE_SOON = 1, /* underlying cred key will expire soon */ RPC_CRED_NOTIFY_TIMEOUT = 2, /* nofity generic cred when underlying key will expire soon */ @@ -82,6 +81,9 @@ struct rpc_cred { #define RPCAUTH_CRED_MAGIC 0x0f4aa4f0 +/* rpc_auth au_flags */ +#define RPCAUTH_AUTH_NO_CRKEY_TIMEOUT 0x0001 /* underlying cred has no key timeout */ + /* * Client authentication handle */ @@ -196,7 +198,7 @@ void rpcauth_destroy_credcache(struct rpc_auth *); void rpcauth_clear_credcache(struct rpc_cred_cache *); int rpcauth_key_timeout_notify(struct rpc_auth *, struct rpc_cred *); -bool rpcauth_cred_key_to_expire(struct rpc_cred *); +bool rpcauth_cred_key_to_expire(struct rpc_auth *, struct rpc_cred *); char * rpcauth_stringify_acceptor(struct rpc_cred *); static inline -- cgit v1.2.3