AcoSail/kernel_optimize_test
8
0
Fork 0
forked from luck/tmp_suning_uos_patched
Code Pull Requests Activity
d6d829d92c
kernel_optimize_test/fs/nfs/nfs4trace.h

2273 lines
63 KiB
C
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 22:07:57 +08:00
/* SPDX-License-Identifier: GPL-2.0 */
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
/*
* Copyright (c) 2013 Trond Myklebust <Trond.Myklebust@netapp.com>
*/
#undef TRACE_SYSTEM
#define TRACE_SYSTEM nfs4
#if !defined(_TRACE_NFS4_H) || defined(TRACE_HEADER_MULTI_READ)
#define _TRACE_NFS4_H
#include <linux/tracepoint.h>
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
TRACE_DEFINE_ENUM(EPERM);
TRACE_DEFINE_ENUM(ENOENT);
TRACE_DEFINE_ENUM(EIO);
TRACE_DEFINE_ENUM(ENXIO);
TRACE_DEFINE_ENUM(EACCES);
TRACE_DEFINE_ENUM(EEXIST);
TRACE_DEFINE_ENUM(EXDEV);
TRACE_DEFINE_ENUM(ENOTDIR);
TRACE_DEFINE_ENUM(EISDIR);
TRACE_DEFINE_ENUM(EFBIG);
TRACE_DEFINE_ENUM(ENOSPC);
TRACE_DEFINE_ENUM(EROFS);
TRACE_DEFINE_ENUM(EMLINK);
TRACE_DEFINE_ENUM(ENAMETOOLONG);
TRACE_DEFINE_ENUM(ENOTEMPTY);
TRACE_DEFINE_ENUM(EDQUOT);
TRACE_DEFINE_ENUM(ESTALE);
TRACE_DEFINE_ENUM(EBADHANDLE);
TRACE_DEFINE_ENUM(EBADCOOKIE);
TRACE_DEFINE_ENUM(ENOTSUPP);
TRACE_DEFINE_ENUM(ETOOSMALL);
TRACE_DEFINE_ENUM(EREMOTEIO);
TRACE_DEFINE_ENUM(EBADTYPE);
TRACE_DEFINE_ENUM(EAGAIN);
TRACE_DEFINE_ENUM(ELOOP);
TRACE_DEFINE_ENUM(EOPNOTSUPP);
TRACE_DEFINE_ENUM(EDEADLK);
TRACE_DEFINE_ENUM(ENOMEM);
TRACE_DEFINE_ENUM(EKEYEXPIRED);
TRACE_DEFINE_ENUM(ETIMEDOUT);
TRACE_DEFINE_ENUM(ERESTARTSYS);
TRACE_DEFINE_ENUM(ECONNREFUSED);
TRACE_DEFINE_ENUM(ECONNRESET);
TRACE_DEFINE_ENUM(ENETUNREACH);
TRACE_DEFINE_ENUM(EHOSTUNREACH);
TRACE_DEFINE_ENUM(EHOSTDOWN);
TRACE_DEFINE_ENUM(EPIPE);
TRACE_DEFINE_ENUM(EPFNOSUPPORT);
TRACE_DEFINE_ENUM(EPROTONOSUPPORT);
TRACE_DEFINE_ENUM(NFS4_OK);
TRACE_DEFINE_ENUM(NFS4ERR_ACCESS);
TRACE_DEFINE_ENUM(NFS4ERR_ATTRNOTSUPP);
TRACE_DEFINE_ENUM(NFS4ERR_ADMIN_REVOKED);
TRACE_DEFINE_ENUM(NFS4ERR_BACK_CHAN_BUSY);
TRACE_DEFINE_ENUM(NFS4ERR_BADCHAR);
TRACE_DEFINE_ENUM(NFS4ERR_BADHANDLE);
TRACE_DEFINE_ENUM(NFS4ERR_BADIOMODE);
TRACE_DEFINE_ENUM(NFS4ERR_BADLAYOUT);
TRACE_DEFINE_ENUM(NFS4ERR_BADLABEL);
TRACE_DEFINE_ENUM(NFS4ERR_BADNAME);
TRACE_DEFINE_ENUM(NFS4ERR_BADOWNER);
TRACE_DEFINE_ENUM(NFS4ERR_BADSESSION);
TRACE_DEFINE_ENUM(NFS4ERR_BADSLOT);
TRACE_DEFINE_ENUM(NFS4ERR_BADTYPE);
TRACE_DEFINE_ENUM(NFS4ERR_BADXDR);
TRACE_DEFINE_ENUM(NFS4ERR_BAD_COOKIE);
TRACE_DEFINE_ENUM(NFS4ERR_BAD_HIGH_SLOT);
TRACE_DEFINE_ENUM(NFS4ERR_BAD_RANGE);
TRACE_DEFINE_ENUM(NFS4ERR_BAD_SEQID);
TRACE_DEFINE_ENUM(NFS4ERR_BAD_SESSION_DIGEST);
TRACE_DEFINE_ENUM(NFS4ERR_BAD_STATEID);
TRACE_DEFINE_ENUM(NFS4ERR_CB_PATH_DOWN);
TRACE_DEFINE_ENUM(NFS4ERR_CLID_INUSE);
TRACE_DEFINE_ENUM(NFS4ERR_CLIENTID_BUSY);
TRACE_DEFINE_ENUM(NFS4ERR_COMPLETE_ALREADY);
TRACE_DEFINE_ENUM(NFS4ERR_CONN_NOT_BOUND_TO_SESSION);
TRACE_DEFINE_ENUM(NFS4ERR_DEADLOCK);
TRACE_DEFINE_ENUM(NFS4ERR_DEADSESSION);
TRACE_DEFINE_ENUM(NFS4ERR_DELAY);
TRACE_DEFINE_ENUM(NFS4ERR_DELEG_ALREADY_WANTED);
TRACE_DEFINE_ENUM(NFS4ERR_DELEG_REVOKED);
TRACE_DEFINE_ENUM(NFS4ERR_DENIED);
TRACE_DEFINE_ENUM(NFS4ERR_DIRDELEG_UNAVAIL);
TRACE_DEFINE_ENUM(NFS4ERR_DQUOT);
TRACE_DEFINE_ENUM(NFS4ERR_ENCR_ALG_UNSUPP);
TRACE_DEFINE_ENUM(NFS4ERR_EXIST);
TRACE_DEFINE_ENUM(NFS4ERR_EXPIRED);
TRACE_DEFINE_ENUM(NFS4ERR_FBIG);
TRACE_DEFINE_ENUM(NFS4ERR_FHEXPIRED);
TRACE_DEFINE_ENUM(NFS4ERR_FILE_OPEN);
TRACE_DEFINE_ENUM(NFS4ERR_GRACE);
TRACE_DEFINE_ENUM(NFS4ERR_HASH_ALG_UNSUPP);
TRACE_DEFINE_ENUM(NFS4ERR_INVAL);
TRACE_DEFINE_ENUM(NFS4ERR_IO);
TRACE_DEFINE_ENUM(NFS4ERR_ISDIR);
TRACE_DEFINE_ENUM(NFS4ERR_LAYOUTTRYLATER);
TRACE_DEFINE_ENUM(NFS4ERR_LAYOUTUNAVAILABLE);
TRACE_DEFINE_ENUM(NFS4ERR_LEASE_MOVED);
TRACE_DEFINE_ENUM(NFS4ERR_LOCKED);
TRACE_DEFINE_ENUM(NFS4ERR_LOCKS_HELD);
TRACE_DEFINE_ENUM(NFS4ERR_LOCK_RANGE);
TRACE_DEFINE_ENUM(NFS4ERR_MINOR_VERS_MISMATCH);
TRACE_DEFINE_ENUM(NFS4ERR_MLINK);
TRACE_DEFINE_ENUM(NFS4ERR_MOVED);
TRACE_DEFINE_ENUM(NFS4ERR_NAMETOOLONG);
TRACE_DEFINE_ENUM(NFS4ERR_NOENT);
TRACE_DEFINE_ENUM(NFS4ERR_NOFILEHANDLE);
TRACE_DEFINE_ENUM(NFS4ERR_NOMATCHING_LAYOUT);
TRACE_DEFINE_ENUM(NFS4ERR_NOSPC);
TRACE_DEFINE_ENUM(NFS4ERR_NOTDIR);
TRACE_DEFINE_ENUM(NFS4ERR_NOTEMPTY);
TRACE_DEFINE_ENUM(NFS4ERR_NOTSUPP);
TRACE_DEFINE_ENUM(NFS4ERR_NOT_ONLY_OP);
TRACE_DEFINE_ENUM(NFS4ERR_NOT_SAME);
TRACE_DEFINE_ENUM(NFS4ERR_NO_GRACE);
TRACE_DEFINE_ENUM(NFS4ERR_NXIO);
TRACE_DEFINE_ENUM(NFS4ERR_OLD_STATEID);
TRACE_DEFINE_ENUM(NFS4ERR_OPENMODE);
TRACE_DEFINE_ENUM(NFS4ERR_OP_ILLEGAL);
TRACE_DEFINE_ENUM(NFS4ERR_OP_NOT_IN_SESSION);
TRACE_DEFINE_ENUM(NFS4ERR_PERM);
TRACE_DEFINE_ENUM(NFS4ERR_PNFS_IO_HOLE);
TRACE_DEFINE_ENUM(NFS4ERR_PNFS_NO_LAYOUT);
TRACE_DEFINE_ENUM(NFS4ERR_RECALLCONFLICT);
TRACE_DEFINE_ENUM(NFS4ERR_RECLAIM_BAD);
TRACE_DEFINE_ENUM(NFS4ERR_RECLAIM_CONFLICT);
TRACE_DEFINE_ENUM(NFS4ERR_REJECT_DELEG);
TRACE_DEFINE_ENUM(NFS4ERR_REP_TOO_BIG);
TRACE_DEFINE_ENUM(NFS4ERR_REP_TOO_BIG_TO_CACHE);
TRACE_DEFINE_ENUM(NFS4ERR_REQ_TOO_BIG);
TRACE_DEFINE_ENUM(NFS4ERR_RESOURCE);
TRACE_DEFINE_ENUM(NFS4ERR_RESTOREFH);
TRACE_DEFINE_ENUM(NFS4ERR_RETRY_UNCACHED_REP);
TRACE_DEFINE_ENUM(NFS4ERR_RETURNCONFLICT);
TRACE_DEFINE_ENUM(NFS4ERR_ROFS);
TRACE_DEFINE_ENUM(NFS4ERR_SAME);
TRACE_DEFINE_ENUM(NFS4ERR_SHARE_DENIED);
TRACE_DEFINE_ENUM(NFS4ERR_SEQUENCE_POS);
TRACE_DEFINE_ENUM(NFS4ERR_SEQ_FALSE_RETRY);
TRACE_DEFINE_ENUM(NFS4ERR_SEQ_MISORDERED);
TRACE_DEFINE_ENUM(NFS4ERR_SERVERFAULT);
TRACE_DEFINE_ENUM(NFS4ERR_STALE);
TRACE_DEFINE_ENUM(NFS4ERR_STALE_CLIENTID);
TRACE_DEFINE_ENUM(NFS4ERR_STALE_STATEID);
TRACE_DEFINE_ENUM(NFS4ERR_SYMLINK);
TRACE_DEFINE_ENUM(NFS4ERR_TOOSMALL);
TRACE_DEFINE_ENUM(NFS4ERR_TOO_MANY_OPS);
TRACE_DEFINE_ENUM(NFS4ERR_UNKNOWN_LAYOUTTYPE);
TRACE_DEFINE_ENUM(NFS4ERR_UNSAFE_COMPOUND);
TRACE_DEFINE_ENUM(NFS4ERR_WRONGSEC);
TRACE_DEFINE_ENUM(NFS4ERR_WRONG_CRED);
TRACE_DEFINE_ENUM(NFS4ERR_WRONG_TYPE);
TRACE_DEFINE_ENUM(NFS4ERR_XDEV);
pNFS/flexfiles: Record resend attempts on I/O failure If the attempt to do pNFS fails, then record what action we take to recover (resend, reset to pnfs or reset to mds). Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:07 +08:00
TRACE_DEFINE_ENUM(NFS4ERR_RESET_TO_MDS);
TRACE_DEFINE_ENUM(NFS4ERR_RESET_TO_PNFS);
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
#define show_nfsv4_errors(error) \
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__print_symbolic(error, \
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
{ NFS4_OK, "OK" }, \
/* Mapped by nfs4_stat_to_errno() */ \
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
{ EPERM, "EPERM" }, \
{ ENOENT, "ENOENT" }, \
{ EIO, "EIO" }, \
{ ENXIO, "ENXIO" }, \
{ EACCES, "EACCES" }, \
{ EEXIST, "EEXIST" }, \
{ EXDEV, "EXDEV" }, \
{ ENOTDIR, "ENOTDIR" }, \
{ EISDIR, "EISDIR" }, \
{ EFBIG, "EFBIG" }, \
{ ENOSPC, "ENOSPC" }, \
{ EROFS, "EROFS" }, \
{ EMLINK, "EMLINK" }, \
{ ENAMETOOLONG, "ENAMETOOLONG" }, \
{ ENOTEMPTY, "ENOTEMPTY" }, \
{ EDQUOT, "EDQUOT" }, \
{ ESTALE, "ESTALE" }, \
{ EBADHANDLE, "EBADHANDLE" }, \
{ EBADCOOKIE, "EBADCOOKIE" }, \
{ ENOTSUPP, "ENOTSUPP" }, \
{ ETOOSMALL, "ETOOSMALL" }, \
{ EREMOTEIO, "EREMOTEIO" }, \
{ EBADTYPE, "EBADTYPE" }, \
{ EAGAIN, "EAGAIN" }, \
{ ELOOP, "ELOOP" }, \
{ EOPNOTSUPP, "EOPNOTSUPP" }, \
{ EDEADLK, "EDEADLK" }, \
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
/* RPC errors */ \
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
{ ENOMEM, "ENOMEM" }, \
{ EKEYEXPIRED, "EKEYEXPIRED" }, \
{ ETIMEDOUT, "ETIMEDOUT" }, \
{ ERESTARTSYS, "ERESTARTSYS" }, \
{ ECONNREFUSED, "ECONNREFUSED" }, \
{ ECONNRESET, "ECONNRESET" }, \
{ ENETUNREACH, "ENETUNREACH" }, \
{ EHOSTUNREACH, "EHOSTUNREACH" }, \
{ EHOSTDOWN, "EHOSTDOWN" }, \
{ EPIPE, "EPIPE" }, \
{ EPFNOSUPPORT, "EPFNOSUPPORT" }, \
{ EPROTONOSUPPORT, "EPROTONOSUPPORT" }, \
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
/* NFSv4 native errors */ \
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
{ NFS4ERR_ACCESS, "ACCESS" }, \
{ NFS4ERR_ATTRNOTSUPP, "ATTRNOTSUPP" }, \
{ NFS4ERR_ADMIN_REVOKED, "ADMIN_REVOKED" }, \
{ NFS4ERR_BACK_CHAN_BUSY, "BACK_CHAN_BUSY" }, \
{ NFS4ERR_BADCHAR, "BADCHAR" }, \
{ NFS4ERR_BADHANDLE, "BADHANDLE" }, \
{ NFS4ERR_BADIOMODE, "BADIOMODE" }, \
{ NFS4ERR_BADLAYOUT, "BADLAYOUT" }, \
{ NFS4ERR_BADLABEL, "BADLABEL" }, \
{ NFS4ERR_BADNAME, "BADNAME" }, \
{ NFS4ERR_BADOWNER, "BADOWNER" }, \
{ NFS4ERR_BADSESSION, "BADSESSION" }, \
{ NFS4ERR_BADSLOT, "BADSLOT" }, \
{ NFS4ERR_BADTYPE, "BADTYPE" }, \
{ NFS4ERR_BADXDR, "BADXDR" }, \
{ NFS4ERR_BAD_COOKIE, "BAD_COOKIE" }, \
{ NFS4ERR_BAD_HIGH_SLOT, "BAD_HIGH_SLOT" }, \
{ NFS4ERR_BAD_RANGE, "BAD_RANGE" }, \
{ NFS4ERR_BAD_SEQID, "BAD_SEQID" }, \
{ NFS4ERR_BAD_SESSION_DIGEST, "BAD_SESSION_DIGEST" }, \
{ NFS4ERR_BAD_STATEID, "BAD_STATEID" }, \
{ NFS4ERR_CB_PATH_DOWN, "CB_PATH_DOWN" }, \
{ NFS4ERR_CLID_INUSE, "CLID_INUSE" }, \
{ NFS4ERR_CLIENTID_BUSY, "CLIENTID_BUSY" }, \
{ NFS4ERR_COMPLETE_ALREADY, "COMPLETE_ALREADY" }, \
{ NFS4ERR_CONN_NOT_BOUND_TO_SESSION, \
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
"CONN_NOT_BOUND_TO_SESSION" }, \
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
{ NFS4ERR_DEADLOCK, "DEADLOCK" }, \
{ NFS4ERR_DEADSESSION, "DEAD_SESSION" }, \
{ NFS4ERR_DELAY, "DELAY" }, \
{ NFS4ERR_DELEG_ALREADY_WANTED, \
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
"DELEG_ALREADY_WANTED" }, \
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
{ NFS4ERR_DELEG_REVOKED, "DELEG_REVOKED" }, \
{ NFS4ERR_DENIED, "DENIED" }, \
{ NFS4ERR_DIRDELEG_UNAVAIL, "DIRDELEG_UNAVAIL" }, \
{ NFS4ERR_DQUOT, "DQUOT" }, \
{ NFS4ERR_ENCR_ALG_UNSUPP, "ENCR_ALG_UNSUPP" }, \
{ NFS4ERR_EXIST, "EXIST" }, \
{ NFS4ERR_EXPIRED, "EXPIRED" }, \
{ NFS4ERR_FBIG, "FBIG" }, \
{ NFS4ERR_FHEXPIRED, "FHEXPIRED" }, \
{ NFS4ERR_FILE_OPEN, "FILE_OPEN" }, \
{ NFS4ERR_GRACE, "GRACE" }, \
{ NFS4ERR_HASH_ALG_UNSUPP, "HASH_ALG_UNSUPP" }, \
{ NFS4ERR_INVAL, "INVAL" }, \
{ NFS4ERR_IO, "IO" }, \
{ NFS4ERR_ISDIR, "ISDIR" }, \
{ NFS4ERR_LAYOUTTRYLATER, "LAYOUTTRYLATER" }, \
{ NFS4ERR_LAYOUTUNAVAILABLE, "LAYOUTUNAVAILABLE" }, \
{ NFS4ERR_LEASE_MOVED, "LEASE_MOVED" }, \
{ NFS4ERR_LOCKED, "LOCKED" }, \
{ NFS4ERR_LOCKS_HELD, "LOCKS_HELD" }, \
{ NFS4ERR_LOCK_RANGE, "LOCK_RANGE" }, \
{ NFS4ERR_MINOR_VERS_MISMATCH, "MINOR_VERS_MISMATCH" }, \
{ NFS4ERR_MLINK, "MLINK" }, \
{ NFS4ERR_MOVED, "MOVED" }, \
{ NFS4ERR_NAMETOOLONG, "NAMETOOLONG" }, \
{ NFS4ERR_NOENT, "NOENT" }, \
{ NFS4ERR_NOFILEHANDLE, "NOFILEHANDLE" }, \
{ NFS4ERR_NOMATCHING_LAYOUT, "NOMATCHING_LAYOUT" }, \
{ NFS4ERR_NOSPC, "NOSPC" }, \
{ NFS4ERR_NOTDIR, "NOTDIR" }, \
{ NFS4ERR_NOTEMPTY, "NOTEMPTY" }, \
{ NFS4ERR_NOTSUPP, "NOTSUPP" }, \
{ NFS4ERR_NOT_ONLY_OP, "NOT_ONLY_OP" }, \
{ NFS4ERR_NOT_SAME, "NOT_SAME" }, \
{ NFS4ERR_NO_GRACE, "NO_GRACE" }, \
{ NFS4ERR_NXIO, "NXIO" }, \
{ NFS4ERR_OLD_STATEID, "OLD_STATEID" }, \
{ NFS4ERR_OPENMODE, "OPENMODE" }, \
{ NFS4ERR_OP_ILLEGAL, "OP_ILLEGAL" }, \
{ NFS4ERR_OP_NOT_IN_SESSION, "OP_NOT_IN_SESSION" }, \
{ NFS4ERR_PERM, "PERM" }, \
{ NFS4ERR_PNFS_IO_HOLE, "PNFS_IO_HOLE" }, \
{ NFS4ERR_PNFS_NO_LAYOUT, "PNFS_NO_LAYOUT" }, \
{ NFS4ERR_RECALLCONFLICT, "RECALLCONFLICT" }, \
{ NFS4ERR_RECLAIM_BAD, "RECLAIM_BAD" }, \
{ NFS4ERR_RECLAIM_CONFLICT, "RECLAIM_CONFLICT" }, \
{ NFS4ERR_REJECT_DELEG, "REJECT_DELEG" }, \
{ NFS4ERR_REP_TOO_BIG, "REP_TOO_BIG" }, \
{ NFS4ERR_REP_TOO_BIG_TO_CACHE, \
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
"REP_TOO_BIG_TO_CACHE" }, \
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
{ NFS4ERR_REQ_TOO_BIG, "REQ_TOO_BIG" }, \
{ NFS4ERR_RESOURCE, "RESOURCE" }, \
{ NFS4ERR_RESTOREFH, "RESTOREFH" }, \
{ NFS4ERR_RETRY_UNCACHED_REP, "RETRY_UNCACHED_REP" }, \
{ NFS4ERR_RETURNCONFLICT, "RETURNCONFLICT" }, \
{ NFS4ERR_ROFS, "ROFS" }, \
{ NFS4ERR_SAME, "SAME" }, \
{ NFS4ERR_SHARE_DENIED, "SHARE_DENIED" }, \
{ NFS4ERR_SEQUENCE_POS, "SEQUENCE_POS" }, \
{ NFS4ERR_SEQ_FALSE_RETRY, "SEQ_FALSE_RETRY" }, \
{ NFS4ERR_SEQ_MISORDERED, "SEQ_MISORDERED" }, \
{ NFS4ERR_SERVERFAULT, "SERVERFAULT" }, \
{ NFS4ERR_STALE, "STALE" }, \
{ NFS4ERR_STALE_CLIENTID, "STALE_CLIENTID" }, \
{ NFS4ERR_STALE_STATEID, "STALE_STATEID" }, \
{ NFS4ERR_SYMLINK, "SYMLINK" }, \
{ NFS4ERR_TOOSMALL, "TOOSMALL" }, \
{ NFS4ERR_TOO_MANY_OPS, "TOO_MANY_OPS" }, \
{ NFS4ERR_UNKNOWN_LAYOUTTYPE, "UNKNOWN_LAYOUTTYPE" }, \
{ NFS4ERR_UNSAFE_COMPOUND, "UNSAFE_COMPOUND" }, \
{ NFS4ERR_WRONGSEC, "WRONGSEC" }, \
{ NFS4ERR_WRONG_CRED, "WRONG_CRED" }, \
{ NFS4ERR_WRONG_TYPE, "WRONG_TYPE" }, \
pNFS/flexfiles: Record resend attempts on I/O failure If the attempt to do pNFS fails, then record what action we take to recover (resend, reset to pnfs or reset to mds). Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:07 +08:00
{ NFS4ERR_XDEV, "XDEV" }, \
/* ***** Internal to Linux NFS client ***** */ \
{ NFS4ERR_RESET_TO_MDS, "RESET_TO_MDS" }, \
{ NFS4ERR_RESET_TO_PNFS, "RESET_TO_PNFS" })
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
#define show_open_flags(flags) \
__print_flags(flags, "|", \
{ O_CREAT, "O_CREAT" }, \
{ O_EXCL, "O_EXCL" }, \
{ O_TRUNC, "O_TRUNC" }, \
{ O_DIRECT, "O_DIRECT" })
#define show_fmode_flags(mode) \
__print_flags(mode, "|", \
{ ((__force unsigned long)FMODE_READ), "READ" }, \
{ ((__force unsigned long)FMODE_WRITE), "WRITE" }, \
{ ((__force unsigned long)FMODE_EXEC), "EXEC" })
NFSv4: Add tracepoints for debugging getattr Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-14 01:01:39 +08:00
#define show_nfs_fattr_flags(valid) \
__print_flags((unsigned long)valid, "|", \
{ NFS_ATTR_FATTR_TYPE, "TYPE" }, \
{ NFS_ATTR_FATTR_MODE, "MODE" }, \
{ NFS_ATTR_FATTR_NLINK, "NLINK" }, \
{ NFS_ATTR_FATTR_OWNER, "OWNER" }, \
{ NFS_ATTR_FATTR_GROUP, "GROUP" }, \
{ NFS_ATTR_FATTR_RDEV, "RDEV" }, \
{ NFS_ATTR_FATTR_SIZE, "SIZE" }, \
{ NFS_ATTR_FATTR_FSID, "FSID" }, \
{ NFS_ATTR_FATTR_FILEID, "FILEID" }, \
{ NFS_ATTR_FATTR_ATIME, "ATIME" }, \
{ NFS_ATTR_FATTR_MTIME, "MTIME" }, \
{ NFS_ATTR_FATTR_CTIME, "CTIME" }, \
{ NFS_ATTR_FATTR_CHANGE, "CHANGE" }, \
{ NFS_ATTR_FATTR_OWNER_NAME, "OWNER_NAME" }, \
{ NFS_ATTR_FATTR_GROUP_NAME, "GROUP_NAME" })
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
DECLARE_EVENT_CLASS(nfs4_clientid_event,
TP_PROTO(
const struct nfs_client *clp,
int error
),
TP_ARGS(clp, error),
TP_STRUCT__entry(
NFS: Avoid RCU usage in tracepoints There isn't an obvious way to acquire and release the RCU lock during a tracepoint, so we can't use the rpc_peeraddr2str() function here. Instead, rely on the client's cl_hostname, which should have similar enough information without needing an rcu_dereference(). Reported-by: Dave Jones <davej@codemonkey.org.uk> Cc: stable@vger.kernel.org # v3.12 Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-02 03:48:43 +08:00
__string(dstaddr, clp->cl_hostname)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
),
TP_fast_assign(
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFS: Avoid RCU usage in tracepoints There isn't an obvious way to acquire and release the RCU lock during a tracepoint, so we can't use the rpc_peeraddr2str() function here. Instead, rely on the client's cl_hostname, which should have similar enough information without needing an rcu_dereference(). Reported-by: Dave Jones <davej@codemonkey.org.uk> Cc: stable@vger.kernel.org # v3.12 Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-02 03:48:43 +08:00
__assign_str(dstaddr, clp->cl_hostname);
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) dstaddr=%s",
-__entry->error,
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
show_nfsv4_errors(__entry->error),
__get_str(dstaddr)
)
);
#define DEFINE_NFS4_CLIENTID_EVENT(name) \
DEFINE_EVENT(nfs4_clientid_event, name, \
TP_PROTO( \
const struct nfs_client *clp, \
int error \
), \
TP_ARGS(clp, error))
DEFINE_NFS4_CLIENTID_EVENT(nfs4_setclientid);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_setclientid_confirm);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_renew);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_renew_async);
#ifdef CONFIG_NFS_V4_1
DEFINE_NFS4_CLIENTID_EVENT(nfs4_exchange_id);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_create_session);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_destroy_session);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_destroy_clientid);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_bind_conn_to_session);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_sequence);
DEFINE_NFS4_CLIENTID_EVENT(nfs4_reclaim_complete);
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
#define show_nfs4_sequence_status_flags(status) \
__print_flags((unsigned long)status, "|", \
{ SEQ4_STATUS_CB_PATH_DOWN, "CB_PATH_DOWN" }, \
{ SEQ4_STATUS_CB_GSS_CONTEXTS_EXPIRING, \
"CB_GSS_CONTEXTS_EXPIRING" }, \
{ SEQ4_STATUS_CB_GSS_CONTEXTS_EXPIRED, \
"CB_GSS_CONTEXTS_EXPIRED" }, \
{ SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED, \
"EXPIRED_ALL_STATE_REVOKED" }, \
{ SEQ4_STATUS_EXPIRED_SOME_STATE_REVOKED, \
"EXPIRED_SOME_STATE_REVOKED" }, \
{ SEQ4_STATUS_ADMIN_STATE_REVOKED, \
"ADMIN_STATE_REVOKED" }, \
{ SEQ4_STATUS_RECALLABLE_STATE_REVOKED, \
"RECALLABLE_STATE_REVOKED" }, \
{ SEQ4_STATUS_LEASE_MOVED, "LEASE_MOVED" }, \
{ SEQ4_STATUS_RESTART_RECLAIM_NEEDED, \
"RESTART_RECLAIM_NEEDED" }, \
{ SEQ4_STATUS_CB_PATH_DOWN_SESSION, \
"CB_PATH_DOWN_SESSION" }, \
{ SEQ4_STATUS_BACKCHANNEL_FAULT, \
"BACKCHANNEL_FAULT" })
TRACE_EVENT(nfs4_sequence_done,
TP_PROTO(
const struct nfs4_session *session,
const struct nfs4_sequence_res *res
),
TP_ARGS(session, res),
TP_STRUCT__entry(
__field(unsigned int, session)
__field(unsigned int, slot_nr)
__field(unsigned int, seq_nr)
__field(unsigned int, highest_slotid)
__field(unsigned int, target_highest_slotid)
__field(unsigned int, status_flags)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
),
TP_fast_assign(
const struct nfs4_slot *sr_slot = res->sr_slot;
__entry->session = nfs_session_id_hash(&session->sess_id);
__entry->slot_nr = sr_slot->slot_nr;
__entry->seq_nr = sr_slot->seq_nr;
__entry->highest_slotid = res->sr_highest_slotid;
__entry->target_highest_slotid =
res->sr_target_highest_slotid;
nfs: fix missing assignment in nfs4_sequence_done tracepoint status_flags not set Signed-off-by: Andrew Elble <aweits@rit.edu> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-02 22:19:43 +08:00
__entry->status_flags = res->sr_status_flags;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = res->sr_status < 0 ?
-res->sr_status : 0;
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) session=0x%08x slot_nr=%u seq_nr=%u "
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
"highest_slotid=%u target_highest_slotid=%u "
"status_flags=%u (%s)",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
show_nfsv4_errors(__entry->error),
__entry->session,
__entry->slot_nr,
__entry->seq_nr,
__entry->highest_slotid,
__entry->target_highest_slotid,
__entry->status_flags,
show_nfs4_sequence_status_flags(__entry->status_flags)
)
);
struct cb_sequenceargs;
struct cb_sequenceres;
TRACE_EVENT(nfs4_cb_sequence,
TP_PROTO(
const struct cb_sequenceargs *args,
const struct cb_sequenceres *res,
__be32 status
),
TP_ARGS(args, res, status),
TP_STRUCT__entry(
__field(unsigned int, session)
__field(unsigned int, slot_nr)
__field(unsigned int, seq_nr)
__field(unsigned int, highest_slotid)
__field(unsigned int, cachethis)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
),
TP_fast_assign(
__entry->session = nfs_session_id_hash(&args->csa_sessionid);
__entry->slot_nr = args->csa_slotid;
__entry->seq_nr = args->csa_sequenceid;
__entry->highest_slotid = args->csa_highestslotid;
__entry->cachethis = args->csa_cachethis;
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = be32_to_cpu(status);
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) session=0x%08x slot_nr=%u seq_nr=%u "
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
"highest_slotid=%u",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4.1: Add tracepoints for debugging slot table operations Add tracepoints to nfs41_setup_sequence and nfs41_sequence_done to track session and slot table state changes. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 05:58:28 +08:00
show_nfsv4_errors(__entry->error),
__entry->session,
__entry->slot_nr,
__entry->seq_nr,
__entry->highest_slotid
)
);
NFS4: Add a trace event to record invalid CB sequence IDs Help debug NFSv4 callback failures. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:47 +08:00
TRACE_EVENT(nfs4_cb_seqid_err,
TP_PROTO(
const struct cb_sequenceargs *args,
__be32 status
),
TP_ARGS(args, status),
TP_STRUCT__entry(
__field(unsigned int, session)
__field(unsigned int, slot_nr)
__field(unsigned int, seq_nr)
__field(unsigned int, highest_slotid)
__field(unsigned int, cachethis)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFS4: Add a trace event to record invalid CB sequence IDs Help debug NFSv4 callback failures. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:47 +08:00
),
TP_fast_assign(
__entry->session = nfs_session_id_hash(&args->csa_sessionid);
__entry->slot_nr = args->csa_slotid;
__entry->seq_nr = args->csa_sequenceid;
__entry->highest_slotid = args->csa_highestslotid;
__entry->cachethis = args->csa_cachethis;
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = be32_to_cpu(status);
NFS4: Add a trace event to record invalid CB sequence IDs Help debug NFSv4 callback failures. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:47 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) session=0x%08x slot_nr=%u seq_nr=%u "
NFS4: Add a trace event to record invalid CB sequence IDs Help debug NFSv4 callback failures. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:47 +08:00
"highest_slotid=%u",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFS4: Add a trace event to record invalid CB sequence IDs Help debug NFSv4 callback failures. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:47 +08:00
show_nfsv4_errors(__entry->error),
__entry->session,
__entry->slot_nr,
__entry->seq_nr,
__entry->highest_slotid
)
);
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
#endif /* CONFIG_NFS_V4_1 */
NFS: Make trace_nfs4_setup_sequence() available to NFS v4.0 This tracepoint displays information about the slot that was chosen for the RPC, in addition to session information. This could be useful information for debugging, and we can set the session id hash to 0 to indicate that there is no session. Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-01-12 02:37:06 +08:00
TRACE_EVENT(nfs4_setup_sequence,
TP_PROTO(
const struct nfs4_session *session,
const struct nfs4_sequence_args *args
),
TP_ARGS(session, args),
TP_STRUCT__entry(
__field(unsigned int, session)
__field(unsigned int, slot_nr)
__field(unsigned int, seq_nr)
__field(unsigned int, highest_used_slotid)
),
TP_fast_assign(
const struct nfs4_slot *sa_slot = args->sa_slot;
__entry->session = session ? nfs_session_id_hash(&session->sess_id) : 0;
__entry->slot_nr = sa_slot->slot_nr;
__entry->seq_nr = sa_slot->seq_nr;
__entry->highest_used_slotid =
sa_slot->table->highest_used_slotid;
),
TP_printk(
"session=0x%08x slot_nr=%u seq_nr=%u "
"highest_used_slotid=%u",
__entry->session,
__entry->slot_nr,
__entry->seq_nr,
__entry->highest_used_slotid
)
);
NFS4: Trace state recovery operation Add a trace point in the main state manager loop to observe state recovery operation. Help track down state recovery bugs. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2019-11-06 00:04:07 +08:00
TRACE_DEFINE_ENUM(NFS4CLNT_MANAGER_RUNNING);
TRACE_DEFINE_ENUM(NFS4CLNT_CHECK_LEASE);
TRACE_DEFINE_ENUM(NFS4CLNT_LEASE_EXPIRED);
TRACE_DEFINE_ENUM(NFS4CLNT_RECLAIM_REBOOT);
TRACE_DEFINE_ENUM(NFS4CLNT_RECLAIM_NOGRACE);
TRACE_DEFINE_ENUM(NFS4CLNT_DELEGRETURN);
TRACE_DEFINE_ENUM(NFS4CLNT_SESSION_RESET);
TRACE_DEFINE_ENUM(NFS4CLNT_LEASE_CONFIRM);
TRACE_DEFINE_ENUM(NFS4CLNT_SERVER_SCOPE_MISMATCH);
TRACE_DEFINE_ENUM(NFS4CLNT_PURGE_STATE);
TRACE_DEFINE_ENUM(NFS4CLNT_BIND_CONN_TO_SESSION);
TRACE_DEFINE_ENUM(NFS4CLNT_MOVED);
TRACE_DEFINE_ENUM(NFS4CLNT_LEASE_MOVED);
TRACE_DEFINE_ENUM(NFS4CLNT_DELEGATION_EXPIRED);
TRACE_DEFINE_ENUM(NFS4CLNT_RUN_MANAGER);
TRACE_DEFINE_ENUM(NFS4CLNT_DELEGRETURN_RUNNING);
#define show_nfs4_clp_state(state) \
__print_flags(state, "|", \
{ NFS4CLNT_MANAGER_RUNNING, "MANAGER_RUNNING" }, \
{ NFS4CLNT_CHECK_LEASE, "CHECK_LEASE" }, \
{ NFS4CLNT_LEASE_EXPIRED, "LEASE_EXPIRED" }, \
{ NFS4CLNT_RECLAIM_REBOOT, "RECLAIM_REBOOT" }, \
{ NFS4CLNT_RECLAIM_NOGRACE, "RECLAIM_NOGRACE" }, \
{ NFS4CLNT_DELEGRETURN, "DELEGRETURN" }, \
{ NFS4CLNT_SESSION_RESET, "SESSION_RESET" }, \
{ NFS4CLNT_LEASE_CONFIRM, "LEASE_CONFIRM" }, \
{ NFS4CLNT_SERVER_SCOPE_MISMATCH, \
"SERVER_SCOPE_MISMATCH" }, \
{ NFS4CLNT_PURGE_STATE, "PURGE_STATE" }, \
{ NFS4CLNT_BIND_CONN_TO_SESSION, \
"BIND_CONN_TO_SESSION" }, \
{ NFS4CLNT_MOVED, "MOVED" }, \
{ NFS4CLNT_LEASE_MOVED, "LEASE_MOVED" }, \
{ NFS4CLNT_DELEGATION_EXPIRED, "DELEGATION_EXPIRED" }, \
{ NFS4CLNT_RUN_MANAGER, "RUN_MANAGER" }, \
{ NFS4CLNT_DELEGRETURN_RUNNING, "DELEGRETURN_RUNNING" })
TRACE_EVENT(nfs4_state_mgr,
TP_PROTO(
const struct nfs_client *clp
),
TP_ARGS(clp),
TP_STRUCT__entry(
__field(unsigned long, state)
__string(hostname, clp->cl_hostname)
),
TP_fast_assign(
__entry->state = clp->cl_state;
__assign_str(hostname, clp->cl_hostname)
),
TP_printk(
"hostname=%s clp state=%s", __get_str(hostname),
show_nfs4_clp_state(__entry->state)
)
)
TRACE_EVENT(nfs4_state_mgr_failed,
TP_PROTO(
const struct nfs_client *clp,
const char *section,
int status
),
TP_ARGS(clp, section, status),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(unsigned long, state)
__string(hostname, clp->cl_hostname)
__string(section, section)
),
TP_fast_assign(
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = status < 0 ? -status : 0;
NFS4: Trace state recovery operation Add a trace point in the main state manager loop to observe state recovery operation. Help track down state recovery bugs. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2019-11-06 00:04:07 +08:00
__entry->state = clp->cl_state;
__assign_str(hostname, clp->cl_hostname);
__assign_str(section, section);
),
TP_printk(
"hostname=%s clp state=%s error=%ld (%s) section=%s",
__get_str(hostname),
show_nfs4_clp_state(__entry->state), -__entry->error,
show_nfsv4_errors(__entry->error), __get_str(section)
)
)
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
TRACE_EVENT(nfs4_xdr_status,
TP_PROTO(
NFS: Record task, client ID, and XID in xdr_status trace points When triggering an nfs_xdr_status trace point, record the task ID and XID of the failing RPC to better pinpoint the problem. This feels like a bit of a layering violation. Suggested-by: Trond Myklebust <trondmy@hammerspace.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:34:09 +08:00
const struct xdr_stream *xdr,
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
u32 op,
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
u32 error
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
),
NFS: Record task, client ID, and XID in xdr_status trace points When triggering an nfs_xdr_status trace point, record the task ID and XID of the failing RPC to better pinpoint the problem. This feels like a bit of a layering violation. Suggested-by: Trond Myklebust <trondmy@hammerspace.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:34:09 +08:00
TP_ARGS(xdr, op, error),
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
TP_STRUCT__entry(
NFS: Record task, client ID, and XID in xdr_status trace points When triggering an nfs_xdr_status trace point, record the task ID and XID of the failing RPC to better pinpoint the problem. This feels like a bit of a layering violation. Suggested-by: Trond Myklebust <trondmy@hammerspace.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:34:09 +08:00
__field(unsigned int, task_id)
__field(unsigned int, client_id)
__field(u32, xid)
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
__field(u32, op)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
),
TP_fast_assign(
NFS: Record task, client ID, and XID in xdr_status trace points When triggering an nfs_xdr_status trace point, record the task ID and XID of the failing RPC to better pinpoint the problem. This feels like a bit of a layering violation. Suggested-by: Trond Myklebust <trondmy@hammerspace.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:34:09 +08:00
const struct rpc_rqst *rqstp = xdr->rqst;
const struct rpc_task *task = rqstp->rq_task;
__entry->task_id = task->tk_pid;
__entry->client_id = task->tk_client->cl_clid;
__entry->xid = be32_to_cpu(rqstp->rq_xid);
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
__entry->op = op;
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = error;
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
),
TP_printk(
NFS: Record task, client ID, and XID in xdr_status trace points When triggering an nfs_xdr_status trace point, record the task ID and XID of the failing RPC to better pinpoint the problem. This feels like a bit of a layering violation. Suggested-by: Trond Myklebust <trondmy@hammerspace.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:34:09 +08:00
"task:%u@%d xid=0x%08x error=%ld (%s) operation=%u",
__entry->task_id, __entry->client_id, __entry->xid,
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error, show_nfsv4_errors(__entry->error),
__entry->op
NFS: Add trace events to report non-zero NFS status codes These can help field troubleshooting without needing the overhead of a full network capture (ie, tcpdump). Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-02-12 00:24:26 +08:00
)
);
NFS4: Report callback authentication errors This seems to be a somewhat common issue with Kerberos NFSv4.0 set-ups. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-12-23 23:28:38 +08:00
DECLARE_EVENT_CLASS(nfs4_cb_error_class,
TP_PROTO(
__be32 xid,
u32 cb_ident
),
TP_ARGS(xid, cb_ident),
TP_STRUCT__entry(
__field(u32, xid)
__field(u32, cbident)
),
TP_fast_assign(
__entry->xid = be32_to_cpu(xid);
__entry->cbident = cb_ident;
),
TP_printk(
"xid=0x%08x cb_ident=0x%08x",
__entry->xid, __entry->cbident
)
);
#define DEFINE_CB_ERROR_EVENT(name) \
DEFINE_EVENT(nfs4_cb_error_class, nfs_cb_##name, \
TP_PROTO( \
__be32 xid, \
u32 cb_ident \
), \
TP_ARGS(xid, cb_ident))
DEFINE_CB_ERROR_EVENT(no_clp);
DEFINE_CB_ERROR_EVENT(badprinc);
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
DECLARE_EVENT_CLASS(nfs4_open_event,
TP_PROTO(
const struct nfs_open_context *ctx,
int flags,
int error
),
TP_ARGS(ctx, flags, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
__field(unsigned int, flags)
__field(unsigned int, fmode)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, dir)
__string(name, ctx->dentry->d_name.name)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
__field(int, openstateid_seq)
__field(u32, openstateid_hash)
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
),
TP_fast_assign(
const struct nfs4_state *state = ctx->state;
const struct inode *inode = NULL;
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = -error;
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
__entry->flags = flags;
__entry->fmode = (__force unsigned int)ctx->mode;
__entry->dev = ctx->dentry->d_sb->s_dev;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
if (!IS_ERR_OR_NULL(state)) {
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
inode = state->inode;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
__entry->openstateid_seq =
be32_to_cpu(state->open_stateid.seqid);
__entry->openstateid_hash =
nfs_stateid_hash(&state->open_stateid);
} else {
__entry->stateid_seq = 0;
__entry->stateid_hash = 0;
__entry->openstateid_seq = 0;
__entry->openstateid_hash = 0;
}
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
if (inode != NULL) {
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
} else {
__entry->fileid = 0;
__entry->fhandle = 0;
}
VFS: normal filesystems (and lustre): d_inode() annotations that's the bulk of filesystem drivers dealing with inodes of their own Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-03-18 06:25:59 +08:00
__entry->dir = NFS_FILEID(d_inode(ctx->dentry->d_parent));
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
__assign_str(name, ctx->dentry->d_name.name);
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) flags=%d (%s) fmode=%s "
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
"fileid=%02x:%02x:%llu fhandle=0x%08x "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"name=%02x:%02x:%llu/%s stateid=%d:0x%08x "
"openstateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
show_nfsv4_errors(__entry->error),
__entry->flags,
show_open_flags(__entry->flags),
show_fmode_flags(__entry->fmode),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__get_str(name),
__entry->stateid_seq, __entry->stateid_hash,
__entry->openstateid_seq, __entry->openstateid_hash
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
)
);
#define DEFINE_NFS4_OPEN_EVENT(name) \
DEFINE_EVENT(nfs4_open_event, name, \
TP_PROTO( \
const struct nfs_open_context *ctx, \
int flags, \
int error \
), \
TP_ARGS(ctx, flags, error))
DEFINE_NFS4_OPEN_EVENT(nfs4_open_reclaim);
DEFINE_NFS4_OPEN_EVENT(nfs4_open_expired);
DEFINE_NFS4_OPEN_EVENT(nfs4_open_file);
Adding tracepoint to cached open Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:42 +08:00
TRACE_EVENT(nfs4_cached_open,
TP_PROTO(
const struct nfs4_state *state
),
TP_ARGS(state),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(unsigned int, fmode)
__field(int, stateid_seq)
__field(u32, stateid_hash)
),
TP_fast_assign(
const struct inode *inode = state->inode;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->fmode = (__force unsigned int)state->state;
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
),
TP_printk(
"fmode=%s fileid=%02x:%02x:%llu "
"fhandle=0x%08x stateid=%d:0x%08x",
__entry->fmode ? show_fmode_flags(__entry->fmode) :
"closed",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash
)
);
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
TRACE_EVENT(nfs4_close,
TP_PROTO(
const struct nfs4_state *state,
const struct nfs_closeargs *args,
const struct nfs_closeres *res,
int error
),
TP_ARGS(state, args, res, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(unsigned int, fmode)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
),
TP_fast_assign(
const struct inode *inode = state->inode;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->fmode = (__force unsigned int)state->state;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(args->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&args->stateid);
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fmode=%s fileid=%02x:%02x:%llu "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"fhandle=0x%08x openstateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
show_nfsv4_errors(__entry->error),
__entry->fmode ? show_fmode_flags(__entry->fmode) :
"closed",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash
NFSv4: Add tracepoints for debugging file open Set up basic tracepoints for debugging NFSv4 file open/close Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:19:27 +08:00
)
);
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
TRACE_DEFINE_ENUM(F_GETLK);
TRACE_DEFINE_ENUM(F_SETLK);
TRACE_DEFINE_ENUM(F_SETLKW);
TRACE_DEFINE_ENUM(F_RDLCK);
TRACE_DEFINE_ENUM(F_WRLCK);
TRACE_DEFINE_ENUM(F_UNLCK);
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
#define show_lock_cmd(type) \
__print_symbolic((int)type, \
{ F_GETLK, "GETLK" }, \
{ F_SETLK, "SETLK" }, \
{ F_SETLKW, "SETLKW" })
#define show_lock_type(type) \
__print_symbolic((int)type, \
{ F_RDLCK, "RDLCK" }, \
{ F_WRLCK, "WRLCK" }, \
{ F_UNLCK, "UNLCK" })
DECLARE_EVENT_CLASS(nfs4_lock_event,
TP_PROTO(
const struct file_lock *request,
const struct nfs4_state *state,
int cmd,
int error
),
TP_ARGS(request, state, cmd, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
__field(int, cmd)
__field(char, type)
__field(loff_t, start)
__field(loff_t, end)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
),
TP_fast_assign(
const struct inode *inode = state->inode;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
__entry->cmd = cmd;
__entry->type = request->fl_type;
__entry->start = request->fl_start;
__entry->end = request->fl_end;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) cmd=%s:%s range=%lld:%lld "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"stateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
show_nfsv4_errors(__entry->error),
show_lock_cmd(__entry->cmd),
show_lock_type(__entry->type),
(long long)__entry->start,
(long long)__entry->end,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash
NFSv4: Add tracepoints for debugging file locking Set up basic tracepoints for debugging NFSv4 file lock/unlock Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:35:20 +08:00
)
);
#define DEFINE_NFS4_LOCK_EVENT(name) \
DEFINE_EVENT(nfs4_lock_event, name, \
TP_PROTO( \
const struct file_lock *request, \
const struct nfs4_state *state, \
int cmd, \
int error \
), \
TP_ARGS(request, state, cmd, error))
DEFINE_NFS4_LOCK_EVENT(nfs4_get_lock);
DEFINE_NFS4_LOCK_EVENT(nfs4_unlock);
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
TRACE_EVENT(nfs4_set_lock,
TP_PROTO(
const struct file_lock *request,
const struct nfs4_state *state,
const nfs4_stateid *lockstateid,
int cmd,
int error
),
TP_ARGS(request, state, lockstateid, cmd, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, cmd)
__field(char, type)
__field(loff_t, start)
__field(loff_t, end)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(int, stateid_seq)
__field(u32, stateid_hash)
__field(int, lockstateid_seq)
__field(u32, lockstateid_hash)
),
TP_fast_assign(
const struct inode *inode = state->inode;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->cmd = cmd;
__entry->type = request->fl_type;
__entry->start = request->fl_start;
__entry->end = request->fl_end;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
__entry->lockstateid_seq =
be32_to_cpu(lockstateid->seqid);
__entry->lockstateid_hash =
nfs_stateid_hash(lockstateid);
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) cmd=%s:%s range=%lld:%lld "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"stateid=%d:0x%08x lockstateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
show_nfsv4_errors(__entry->error),
show_lock_cmd(__entry->cmd),
show_lock_type(__entry->type),
(long long)__entry->start,
(long long)__entry->end,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash,
__entry->lockstateid_seq, __entry->lockstateid_hash
)
);
NFS4: Trace lock reclaims One of the most frustrating messages our sustaining team sees is the "Lock reclaim failed!" message. Add some observability in the client's lock reclaim logic so we can capture better data the first time a problem occurs. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2019-11-06 00:04:13 +08:00
TRACE_DEFINE_ENUM(LK_STATE_IN_USE);
TRACE_DEFINE_ENUM(NFS_DELEGATED_STATE);
TRACE_DEFINE_ENUM(NFS_OPEN_STATE);
TRACE_DEFINE_ENUM(NFS_O_RDONLY_STATE);
TRACE_DEFINE_ENUM(NFS_O_WRONLY_STATE);
TRACE_DEFINE_ENUM(NFS_O_RDWR_STATE);
TRACE_DEFINE_ENUM(NFS_STATE_RECLAIM_REBOOT);
TRACE_DEFINE_ENUM(NFS_STATE_RECLAIM_NOGRACE);
TRACE_DEFINE_ENUM(NFS_STATE_POSIX_LOCKS);
TRACE_DEFINE_ENUM(NFS_STATE_RECOVERY_FAILED);
TRACE_DEFINE_ENUM(NFS_STATE_MAY_NOTIFY_LOCK);
TRACE_DEFINE_ENUM(NFS_STATE_CHANGE_WAIT);
TRACE_DEFINE_ENUM(NFS_CLNT_DST_SSC_COPY_STATE);
TRACE_DEFINE_ENUM(NFS_CLNT_SRC_SSC_COPY_STATE);
TRACE_DEFINE_ENUM(NFS_SRV_SSC_COPY_STATE);
#define show_nfs4_state_flags(flags) \
__print_flags(flags, "|", \
{ LK_STATE_IN_USE, "IN_USE" }, \
{ NFS_DELEGATED_STATE, "DELEGATED" }, \
{ NFS_OPEN_STATE, "OPEN" }, \
{ NFS_O_RDONLY_STATE, "O_RDONLY" }, \
{ NFS_O_WRONLY_STATE, "O_WRONLY" }, \
{ NFS_O_RDWR_STATE, "O_RDWR" }, \
{ NFS_STATE_RECLAIM_REBOOT, "RECLAIM_REBOOT" }, \
{ NFS_STATE_RECLAIM_NOGRACE, "RECLAIM_NOGRACE" }, \
{ NFS_STATE_POSIX_LOCKS, "POSIX_LOCKS" }, \
{ NFS_STATE_RECOVERY_FAILED, "RECOVERY_FAILED" }, \
{ NFS_STATE_MAY_NOTIFY_LOCK, "MAY_NOTIFY_LOCK" }, \
{ NFS_STATE_CHANGE_WAIT, "CHANGE_WAIT" }, \
{ NFS_CLNT_DST_SSC_COPY_STATE, "CLNT_DST_SSC_COPY" }, \
{ NFS_CLNT_SRC_SSC_COPY_STATE, "CLNT_SRC_SSC_COPY" }, \
{ NFS_SRV_SSC_COPY_STATE, "SRV_SSC_COPY" })
#define show_nfs4_lock_flags(flags) \
__print_flags(flags, "|", \
{ BIT(NFS_LOCK_INITIALIZED), "INITIALIZED" }, \
{ BIT(NFS_LOCK_LOST), "LOST" })
TRACE_EVENT(nfs4_state_lock_reclaim,
TP_PROTO(
const struct nfs4_state *state,
const struct nfs4_lock_state *lock
),
TP_ARGS(state, lock),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(unsigned long, state_flags)
__field(unsigned long, lock_flags)
__field(int, stateid_seq)
__field(u32, stateid_hash)
),
TP_fast_assign(
const struct inode *inode = state->inode;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->state_flags = state->flags;
__entry->lock_flags = lock->ls_flags;
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"stateid=%d:0x%08x state_flags=%s lock_flags=%s",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid, __entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash,
show_nfs4_state_flags(__entry->state_flags),
show_nfs4_lock_flags(__entry->lock_flags)
)
)
NFSv4: Add tracepoints for debugging delegations Set up tracepoints to track when delegations are set, reclaimed, returned by the client, or recalled by the server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 22:36:56 +08:00
DECLARE_EVENT_CLASS(nfs4_set_delegation_event,
TP_PROTO(
const struct inode *inode,
fmode_t fmode
),
TP_ARGS(inode, fmode),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(unsigned int, fmode)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->fmode = (__force unsigned int)fmode;
),
TP_printk(
"fmode=%s fileid=%02x:%02x:%llu fhandle=0x%08x",
show_fmode_flags(__entry->fmode),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle
)
);
#define DEFINE_NFS4_SET_DELEGATION_EVENT(name) \
DEFINE_EVENT(nfs4_set_delegation_event, name, \
TP_PROTO( \
const struct inode *inode, \
fmode_t fmode \
), \
TP_ARGS(inode, fmode))
DEFINE_NFS4_SET_DELEGATION_EVENT(nfs4_set_delegation);
DEFINE_NFS4_SET_DELEGATION_EVENT(nfs4_reclaim_delegation);
TRACE_EVENT(nfs4_delegreturn_exit,
TP_PROTO(
const struct nfs4_delegreturnargs *args,
const struct nfs4_delegreturnres *res,
int error
),
TP_ARGS(args, res, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
NFSv4: Add tracepoints for debugging delegations Set up tracepoints to track when delegations are set, reclaimed, returned by the client, or recalled by the server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 22:36:56 +08:00
),
TP_fast_assign(
__entry->dev = res->server->s_dev;
__entry->fhandle = nfs_fhandle_hash(args->fhandle);
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(args->stateid->seqid);
__entry->stateid_hash =
nfs_stateid_hash(args->stateid);
NFSv4: Add tracepoints for debugging delegations Set up tracepoints to track when delegations are set, reclaimed, returned by the client, or recalled by the server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 22:36:56 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) dev=%02x:%02x fhandle=0x%08x "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"stateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging delegations Set up tracepoints to track when delegations are set, reclaimed, returned by the client, or recalled by the server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 22:36:56 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash
NFSv4: Add tracepoints for debugging delegations Set up tracepoints to track when delegations are set, reclaimed, returned by the client, or recalled by the server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 22:36:56 +08:00
)
);
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
#ifdef CONFIG_NFS_V4_1
DECLARE_EVENT_CLASS(nfs4_test_stateid_event,
TP_PROTO(
const struct nfs4_state *state,
const struct nfs4_lock_state *lsp,
int error
),
TP_ARGS(state, lsp, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
),
TP_fast_assign(
const struct inode *inode = state->inode;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"stateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash
NFSv4.1: Add tracepoints for debugging test_stateid events Add tracepoints to detect issues with the TEST_STATEID operation. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-21 09:59:40 +08:00
)
);
#define DEFINE_NFS4_TEST_STATEID_EVENT(name) \
DEFINE_EVENT(nfs4_test_stateid_event, name, \
TP_PROTO( \
const struct nfs4_state *state, \
const struct nfs4_lock_state *lsp, \
int error \
), \
TP_ARGS(state, lsp, error))
DEFINE_NFS4_TEST_STATEID_EVENT(nfs4_test_delegation_stateid);
DEFINE_NFS4_TEST_STATEID_EVENT(nfs4_test_open_stateid);
DEFINE_NFS4_TEST_STATEID_EVENT(nfs4_test_lock_stateid);
#endif /* CONFIG_NFS_V4_1 */
NFSv4: Add tracepoints for debugging lookup/create operations Set up basic tracepoints for debugging NFSv4 lookup, unlink/remove, symlink, mkdir, mknod, fs_locations and secinfo. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:45:55 +08:00
DECLARE_EVENT_CLASS(nfs4_lookup_event,
TP_PROTO(
const struct inode *dir,
const struct qstr *name,
int error
),
TP_ARGS(dir, name, error),
TP_STRUCT__entry(
__field(dev_t, dev)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging lookup/create operations Set up basic tracepoints for debugging NFSv4 lookup, unlink/remove, symlink, mkdir, mknod, fs_locations and secinfo. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:45:55 +08:00
__field(u64, dir)
__string(name, name->name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = -error;
NFSv4: Add tracepoints for debugging lookup/create operations Set up basic tracepoints for debugging NFSv4 lookup, unlink/remove, symlink, mkdir, mknod, fs_locations and secinfo. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:45:55 +08:00
__assign_str(name, name->name);
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) name=%02x:%02x:%llu/%s",
-__entry->error,
NFSv4: Add tracepoints for debugging lookup/create operations Set up basic tracepoints for debugging NFSv4 lookup, unlink/remove, symlink, mkdir, mknod, fs_locations and secinfo. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:45:55 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
#define DEFINE_NFS4_LOOKUP_EVENT(name) \
DEFINE_EVENT(nfs4_lookup_event, name, \
TP_PROTO( \
const struct inode *dir, \
const struct qstr *name, \
int error \
), \
TP_ARGS(dir, name, error))
DEFINE_NFS4_LOOKUP_EVENT(nfs4_lookup);
DEFINE_NFS4_LOOKUP_EVENT(nfs4_symlink);
DEFINE_NFS4_LOOKUP_EVENT(nfs4_mkdir);
DEFINE_NFS4_LOOKUP_EVENT(nfs4_mknod);
DEFINE_NFS4_LOOKUP_EVENT(nfs4_remove);
DEFINE_NFS4_LOOKUP_EVENT(nfs4_get_fs_locations);
DEFINE_NFS4_LOOKUP_EVENT(nfs4_secinfo);
nfs4: add NFSv4 LOOKUPP handlers This will be needed in order to implement the get_parent export op for nfsd. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-06-29 21:34:52 +08:00
TRACE_EVENT(nfs4_lookupp,
TP_PROTO(
const struct inode *inode,
int error
),
TP_ARGS(inode, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, ino)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
nfs4: add NFSv4 LOOKUPP handlers This will be needed in order to implement the get_parent export op for nfsd. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-06-29 21:34:52 +08:00
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->ino = NFS_FILEID(inode);
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
nfs4: add NFSv4 LOOKUPP handlers This will be needed in order to implement the get_parent export op for nfsd. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-06-29 21:34:52 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) inode=%02x:%02x:%llu",
-__entry->error,
nfs4: add NFSv4 LOOKUPP handlers This will be needed in order to implement the get_parent export op for nfsd. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-06-29 21:34:52 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->ino
)
);
NFSv4: Add tracepoints for debugging rename Add tracepoints to debug renames. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 05:08:26 +08:00
TRACE_EVENT(nfs4_rename,
TP_PROTO(
const struct inode *olddir,
const struct qstr *oldname,
const struct inode *newdir,
const struct qstr *newname,
int error
),
TP_ARGS(olddir, oldname, newdir, newname, error),
TP_STRUCT__entry(
__field(dev_t, dev)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging rename Add tracepoints to debug renames. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 05:08:26 +08:00
__field(u64, olddir)
__string(oldname, oldname->name)
__field(u64, newdir)
__string(newname, newname->name)
),
TP_fast_assign(
__entry->dev = olddir->i_sb->s_dev;
__entry->olddir = NFS_FILEID(olddir);
__entry->newdir = NFS_FILEID(newdir);
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: Add tracepoints for debugging rename Add tracepoints to debug renames. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 05:08:26 +08:00
__assign_str(oldname, oldname->name);
__assign_str(newname, newname->name);
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) oldname=%02x:%02x:%llu/%s "
NFSv4: Add tracepoints for debugging rename Add tracepoints to debug renames. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 05:08:26 +08:00
"newname=%02x:%02x:%llu/%s",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging rename Add tracepoints to debug renames. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 05:08:26 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->olddir,
__get_str(oldname),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->newdir,
__get_str(newname)
)
);
NFSv4: Add tracepoints for debugging inode manipulations Set up basic tracepoints for debugging NFSv4 setattr, access, readlink, readdir, get_acl set_acl get_security_label, and set_security_label. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:58:42 +08:00
DECLARE_EVENT_CLASS(nfs4_inode_event,
TP_PROTO(
const struct inode *inode,
int error
),
TP_ARGS(inode, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging inode manipulations Set up basic tracepoints for debugging NFSv4 setattr, access, readlink, readdir, get_acl set_acl get_security_label, and set_security_label. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:58:42 +08:00
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: Add tracepoints for debugging inode manipulations Set up basic tracepoints for debugging NFSv4 setattr, access, readlink, readdir, get_acl set_acl get_security_label, and set_security_label. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:58:42 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x",
-__entry->error,
NFSv4: Add tracepoints for debugging inode manipulations Set up basic tracepoints for debugging NFSv4 setattr, access, readlink, readdir, get_acl set_acl get_security_label, and set_security_label. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:58:42 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle
)
);
#define DEFINE_NFS4_INODE_EVENT(name) \
DEFINE_EVENT(nfs4_inode_event, name, \
TP_PROTO( \
const struct inode *inode, \
int error \
), \
TP_ARGS(inode, error))
DEFINE_NFS4_INODE_EVENT(nfs4_access);
DEFINE_NFS4_INODE_EVENT(nfs4_readlink);
DEFINE_NFS4_INODE_EVENT(nfs4_readdir);
DEFINE_NFS4_INODE_EVENT(nfs4_get_acl);
DEFINE_NFS4_INODE_EVENT(nfs4_set_acl);
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
DEFINE_NFS4_INODE_EVENT(nfs4_get_security_label);
DEFINE_NFS4_INODE_EVENT(nfs4_set_security_label);
#endif /* CONFIG_NFS_V4_SECURITY_LABEL */
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
DECLARE_EVENT_CLASS(nfs4_inode_stateid_event,
TP_PROTO(
const struct inode *inode,
const nfs4_stateid *stateid,
int error
),
TP_ARGS(inode, stateid, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(stateid->seqid);
__entry->stateid_hash =
nfs_stateid_hash(stateid);
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"stateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash
)
);
#define DEFINE_NFS4_INODE_STATEID_EVENT(name) \
DEFINE_EVENT(nfs4_inode_stateid_event, name, \
TP_PROTO( \
const struct inode *inode, \
const nfs4_stateid *stateid, \
int error \
), \
TP_ARGS(inode, stateid, error))
DEFINE_NFS4_INODE_STATEID_EVENT(nfs4_setattr);
DEFINE_NFS4_INODE_STATEID_EVENT(nfs4_delegreturn);
NFSv4: Add a tracepoint to document open stateid updates Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-07 04:28:02 +08:00
DEFINE_NFS4_INODE_STATEID_EVENT(nfs4_open_stateid_update);
DEFINE_NFS4_INODE_STATEID_EVENT(nfs4_open_stateid_update_wait);
NFSv4: Add tracepoints for debugging inode manipulations Set up basic tracepoints for debugging NFSv4 setattr, access, readlink, readdir, get_acl set_acl get_security_label, and set_security_label. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 04:58:42 +08:00
NFSv4: Add tracepoints for debugging getattr Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-14 01:01:39 +08:00
DECLARE_EVENT_CLASS(nfs4_getattr_event,
TP_PROTO(
const struct nfs_server *server,
const struct nfs_fh *fhandle,
const struct nfs_fattr *fattr,
int error
),
TP_ARGS(server, fhandle, fattr, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(unsigned int, valid)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging getattr Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-14 01:01:39 +08:00
),
TP_fast_assign(
__entry->dev = server->s_dev;
__entry->valid = fattr->valid;
__entry->fhandle = nfs_fhandle_hash(fhandle);
__entry->fileid = (fattr->valid & NFS_ATTR_FATTR_FILEID) ? fattr->fileid : 0;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: Add tracepoints for debugging getattr Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-14 01:01:39 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
NFSv4: Add tracepoints for debugging getattr Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-14 01:01:39 +08:00
"valid=%s",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging getattr Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-14 01:01:39 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
show_nfs_fattr_flags(__entry->valid)
)
);
#define DEFINE_NFS4_GETATTR_EVENT(name) \
DEFINE_EVENT(nfs4_getattr_event, name, \
TP_PROTO( \
const struct nfs_server *server, \
const struct nfs_fh *fhandle, \
const struct nfs_fattr *fattr, \
int error \
), \
TP_ARGS(server, fhandle, fattr, error))
DEFINE_NFS4_GETATTR_EVENT(nfs4_getattr);
DEFINE_NFS4_GETATTR_EVENT(nfs4_lookup_root);
DEFINE_NFS4_GETATTR_EVENT(nfs4_fsinfo);
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
DECLARE_EVENT_CLASS(nfs4_inode_callback_event,
TP_PROTO(
const struct nfs_client *clp,
const struct nfs_fh *fhandle,
const struct inode *inode,
int error
),
TP_ARGS(clp, fhandle, inode, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
NFS: Avoid RCU usage in tracepoints There isn't an obvious way to acquire and release the RCU lock during a tracepoint, so we can't use the rpc_peeraddr2str() function here. Instead, rely on the client's cl_hostname, which should have similar enough information without needing an rcu_dereference(). Reported-by: Dave Jones <davej@codemonkey.org.uk> Cc: stable@vger.kernel.org # v3.12 Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-02 03:48:43 +08:00
__string(dstaddr, clp ? clp->cl_hostname : "unknown")
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
),
TP_fast_assign(
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
__entry->fhandle = nfs_fhandle_hash(fhandle);
NFSv4: Fix a tracepoint Oops in initiate_file_draining() Now that the value of 'ino' can be NULL or an ERR_PTR(), we need to change the test in the tracepoint. Fixes: ce5624f7e6675 ("NFSv4: Return NFS4ERR_DELAY when a layout fails...") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: stable@vger.kernel.org # v4.17+ Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-09-06 02:07:13 +08:00
if (!IS_ERR_OR_NULL(inode)) {
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
__entry->fileid = NFS_FILEID(inode);
__entry->dev = inode->i_sb->s_dev;
} else {
__entry->fileid = 0;
__entry->dev = 0;
}
NFS: Avoid RCU usage in tracepoints There isn't an obvious way to acquire and release the RCU lock during a tracepoint, so we can't use the rpc_peeraddr2str() function here. Instead, rely on the client's cl_hostname, which should have similar enough information without needing an rcu_dereference(). Reported-by: Dave Jones <davej@codemonkey.org.uk> Cc: stable@vger.kernel.org # v3.12 Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-02 03:48:43 +08:00
__assign_str(dstaddr, clp ? clp->cl_hostname : "unknown")
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
"dstaddr=%s",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__get_str(dstaddr)
)
);
#define DEFINE_NFS4_INODE_CALLBACK_EVENT(name) \
DEFINE_EVENT(nfs4_inode_callback_event, name, \
TP_PROTO( \
const struct nfs_client *clp, \
const struct nfs_fh *fhandle, \
const struct inode *inode, \
int error \
), \
TP_ARGS(clp, fhandle, inode, error))
DEFINE_NFS4_INODE_CALLBACK_EVENT(nfs4_cb_getattr);
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
DECLARE_EVENT_CLASS(nfs4_inode_stateid_callback_event,
TP_PROTO(
const struct nfs_client *clp,
const struct nfs_fh *fhandle,
const struct inode *inode,
const nfs4_stateid *stateid,
int error
),
TP_ARGS(clp, fhandle, inode, stateid, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
NFS: Avoid RCU usage in tracepoints There isn't an obvious way to acquire and release the RCU lock during a tracepoint, so we can't use the rpc_peeraddr2str() function here. Instead, rely on the client's cl_hostname, which should have similar enough information without needing an rcu_dereference(). Reported-by: Dave Jones <davej@codemonkey.org.uk> Cc: stable@vger.kernel.org # v3.12 Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-02 03:48:43 +08:00
__string(dstaddr, clp ? clp->cl_hostname : "unknown")
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
),
TP_fast_assign(
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
__entry->fhandle = nfs_fhandle_hash(fhandle);
NFSv4: Fix a tracepoint Oops in initiate_file_draining() Now that the value of 'ino' can be NULL or an ERR_PTR(), we need to change the test in the tracepoint. Fixes: ce5624f7e6675 ("NFSv4: Return NFS4ERR_DELAY when a layout fails...") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: stable@vger.kernel.org # v4.17+ Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-08-23 23:02:49 +08:00
if (!IS_ERR_OR_NULL(inode)) {
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
__entry->fileid = NFS_FILEID(inode);
__entry->dev = inode->i_sb->s_dev;
} else {
__entry->fileid = 0;
__entry->dev = 0;
}
NFS: Avoid RCU usage in tracepoints There isn't an obvious way to acquire and release the RCU lock during a tracepoint, so we can't use the rpc_peeraddr2str() function here. Instead, rely on the client's cl_hostname, which should have similar enough information without needing an rcu_dereference(). Reported-by: Dave Jones <davej@codemonkey.org.uk> Cc: stable@vger.kernel.org # v3.12 Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-11-02 03:48:43 +08:00
__assign_str(dstaddr, clp ? clp->cl_hostname : "unknown")
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
__entry->stateid_seq =
be32_to_cpu(stateid->seqid);
__entry->stateid_hash =
nfs_stateid_hash(stateid);
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
"stateid=%d:0x%08x dstaddr=%s",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: List stateid information in the callback tracepoints The stateid is extremely valuable when debugging. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-29 03:07:23 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->stateid_seq, __entry->stateid_hash,
__get_str(dstaddr)
)
);
#define DEFINE_NFS4_INODE_STATEID_CALLBACK_EVENT(name) \
DEFINE_EVENT(nfs4_inode_stateid_callback_event, name, \
TP_PROTO( \
const struct nfs_client *clp, \
const struct nfs_fh *fhandle, \
const struct inode *inode, \
const nfs4_stateid *stateid, \
int error \
), \
TP_ARGS(clp, fhandle, inode, stateid, error))
DEFINE_NFS4_INODE_STATEID_CALLBACK_EVENT(nfs4_cb_recall);
DEFINE_NFS4_INODE_STATEID_CALLBACK_EVENT(nfs4_cb_layoutrecall_file);
NFSv4: Add a tracepoint for CB_GETATTR Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 09:07:54 +08:00
NFSv4: Add tracepoints for debugging the idmapper Add tracepoints to help debug uid/gid mappings to username/group. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 23:34:01 +08:00
DECLARE_EVENT_CLASS(nfs4_idmap_event,
TP_PROTO(
const char *name,
int len,
u32 id,
int error
),
TP_ARGS(name, len, id, error),
TP_STRUCT__entry(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Add tracepoints for debugging the idmapper Add tracepoints to help debug uid/gid mappings to username/group. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 23:34:01 +08:00
__field(u32, id)
__dynamic_array(char, name, len > 0 ? len + 1 : 1)
),
TP_fast_assign(
if (len < 0)
len = 0;
__entry->error = error < 0 ? error : 0;
__entry->id = id;
tracing: Use __get_str() when manipulating strings Use __get_str(str) rather than __get_dynamic_array(str) when deadling with strings. It is just a code cleanup, no changes on tracepoint ABI. Link: http://lkml.kernel.org/r/ea260df91817411cca2a1f3db2abd88860094788.1467407618.git.bristot@redhat.com Cc: Trond Myklebust <trond.myklebust@primarydata.com> Cc: Anna Schumaker <anna.schumaker@netapp.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: linux-nfs@vger.kernel.org Suggested-by: Steven Rostedt <rostedt@goodmis.org> Reviewed-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Daniel Bristot de Oliveira <bristot@redhat.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-07-02 07:44:36 +08:00
memcpy(__get_str(name), name, len);
__get_str(name)[len] = 0;
NFSv4: Add tracepoints for debugging the idmapper Add tracepoints to help debug uid/gid mappings to username/group. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 23:34:01 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) id=%u name=%s",
-__entry->error, show_nfsv4_errors(__entry->error),
NFSv4: Add tracepoints for debugging the idmapper Add tracepoints to help debug uid/gid mappings to username/group. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-13 23:34:01 +08:00
__entry->id,
__get_str(name)
)
);
#define DEFINE_NFS4_IDMAP_EVENT(name) \
DEFINE_EVENT(nfs4_idmap_event, name, \
TP_PROTO( \
const char *name, \
int len, \
u32 id, \
int error \
), \
TP_ARGS(name, len, id, error))
DEFINE_NFS4_IDMAP_EVENT(nfs4_map_name_to_uid);
DEFINE_NFS4_IDMAP_EVENT(nfs4_map_group_to_gid);
DEFINE_NFS4_IDMAP_EVENT(nfs4_map_uid_to_name);
DEFINE_NFS4_IDMAP_EVENT(nfs4_map_gid_to_group);
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
DECLARE_EVENT_CLASS(nfs4_read_event,
TP_PROTO(
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
const struct nfs_pgio_header *hdr,
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
int error
),
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
TP_ARGS(hdr, error),
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__field(u32, arg_count)
__field(u32, res_count)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
),
TP_fast_assign(
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
const struct inode *inode = hdr->inode;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
const struct nfs4_state *state =
hdr->args.context->state;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
__entry->dev = inode->i_sb->s_dev;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
__entry->offset = hdr->args.offset;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->arg_count = hdr->args.count;
__entry->res_count = hdr->res.count;
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
"offset=%lld count=%u res=%u stateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset,
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->arg_count, __entry->res_count,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq, __entry->stateid_hash
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
)
);
#define DEFINE_NFS4_READ_EVENT(name) \
DEFINE_EVENT(nfs4_read_event, name, \
TP_PROTO( \
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
const struct nfs_pgio_header *hdr, \
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
int error \
), \
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
TP_ARGS(hdr, error))
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
DEFINE_NFS4_READ_EVENT(nfs4_read);
#ifdef CONFIG_NFS_V4_1
DEFINE_NFS4_READ_EVENT(nfs4_pnfs_read);
#endif /* CONFIG_NFS_V4_1 */
DECLARE_EVENT_CLASS(nfs4_write_event,
TP_PROTO(
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
const struct nfs_pgio_header *hdr,
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
int error
),
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
TP_ARGS(hdr, error),
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__field(u32, arg_count)
__field(u32, res_count)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
),
TP_fast_assign(
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
const struct inode *inode = hdr->inode;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
const struct nfs4_state *state =
hdr->args.context->state;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
__entry->dev = inode->i_sb->s_dev;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
__entry->offset = hdr->args.offset;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->arg_count = hdr->args.count;
__entry->res_count = hdr->res.count;
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
"offset=%lld count=%u res=%u stateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset,
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->arg_count, __entry->res_count,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq, __entry->stateid_hash
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
)
);
#define DEFINE_NFS4_WRITE_EVENT(name) \
DEFINE_EVENT(nfs4_write_event, name, \
TP_PROTO( \
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
const struct nfs_pgio_header *hdr, \
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
int error \
), \
nfs: merge nfs_pgio_data into _header struct nfs_pgio_data only exists as a member of nfs_pgio_header, but is passed around everywhere, because there used to be multiple _data structs per _header. Many of these functions then use the _data to find a pointer to the _header. This patch cleans this up by merging the nfs_pgio_data structure into nfs_pgio_header and passing nfs_pgio_header around instead. Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-06-09 23:48:35 +08:00
TP_ARGS(hdr, error))
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
DEFINE_NFS4_WRITE_EVENT(nfs4_write);
#ifdef CONFIG_NFS_V4_1
DEFINE_NFS4_WRITE_EVENT(nfs4_pnfs_write);
#endif /* CONFIG_NFS_V4_1 */
DECLARE_EVENT_CLASS(nfs4_commit_event,
TP_PROTO(
const struct nfs_commit_data *data,
int error
),
TP_ARGS(data, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__field(loff_t, offset)
__field(u32, count)
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
),
TP_fast_assign(
const struct inode *inode = data->inode;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = data->args.fh ?
data->args.fh : &nfsi->fh;
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
__entry->dev = inode->i_sb->s_dev;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
__entry->offset = data->args.offset;
__entry->count = data->args.count;
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
__entry->error = error < 0 ? -error : 0;
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
NFSv4: Improve read/write/commit tracing Ensure we always return the number of bytes read/written. Also display the pnfs filehandle if it is in use. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:05 +08:00
"offset=%lld count=%u",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset,
__entry->count
)
);
#define DEFINE_NFS4_COMMIT_EVENT(name) \
DEFINE_EVENT(nfs4_commit_event, name, \
TP_PROTO( \
const struct nfs_commit_data *data, \
int error \
), \
TP_ARGS(data, error))
DEFINE_NFS4_COMMIT_EVENT(nfs4_commit);
#ifdef CONFIG_NFS_V4_1
DEFINE_NFS4_COMMIT_EVENT(nfs4_pnfs_commit_ds);
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
TRACE_DEFINE_ENUM(IOMODE_READ);
TRACE_DEFINE_ENUM(IOMODE_RW);
TRACE_DEFINE_ENUM(IOMODE_ANY);
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
#define show_pnfs_iomode(iomode) \
__print_symbolic(iomode, \
{ IOMODE_READ, "READ" }, \
{ IOMODE_RW, "RW" }, \
{ IOMODE_ANY, "ANY" })
TRACE_EVENT(nfs4_layoutget,
TP_PROTO(
const struct nfs_open_context *ctx,
const struct pnfs_layout_range *args,
const struct pnfs_layout_range *res,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
const nfs4_stateid *layout_stateid,
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
int error
),
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
TP_ARGS(ctx, args, res, layout_stateid, error),
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u32, iomode)
__field(u64, offset)
__field(u64, count)
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
__field(unsigned long, error)
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__field(int, stateid_seq)
__field(u32, stateid_hash)
__field(int, layoutstateid_seq)
__field(u32, layoutstateid_hash)
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
),
TP_fast_assign(
VFS: normal filesystems (and lustre): d_inode() annotations that's the bulk of filesystem drivers dealing with inodes of their own Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2015-03-18 06:25:59 +08:00
const struct inode *inode = d_inode(ctx->dentry);
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
const struct nfs4_state *state = ctx->state;
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->iomode = args->iomode;
__entry->offset = args->offset;
__entry->count = args->length;
NFS: Fix fix of show_nfs_errors Casting a negative value to an unsigned long is not the same as converting it to its absolute value. Fixes: 96650e2effa2 ("NFS: Fix show_nfs_errors macros again") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:06 +08:00
__entry->error = error < 0 ? -error : 0;
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
__entry->stateid_seq =
be32_to_cpu(state->stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&state->stateid);
if (!error) {
__entry->layoutstateid_seq =
be32_to_cpu(layout_stateid->seqid);
__entry->layoutstateid_hash =
nfs_stateid_hash(layout_stateid);
} else {
__entry->layoutstateid_seq = 0;
__entry->layoutstateid_hash = 0;
}
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
),
TP_printk(
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
"iomode=%s offset=%llu count=%llu stateid=%d:0x%08x "
"layoutstateid=%d:0x%08x",
NFS: Fix show_nfs_errors macros again I noticed that NFS status values stopped working again. trace_print_symbols_seq() takes an unsigned long. Passing a negative errno or negative NFSERR value just confuses it, and since we're using C macros here and not static inline functions, all bets are off due to implicit type conversion. Straight-line the calling conventions so that error codes are stored in the trace record as positive values in an unsigned long field, mapped to symbolic as an unsigned long, and displayed as a negative value, to continue to enable grepping on "error=-". It's often the case that an error value that is positive is a byte count but when it's negative, it's an error (e.g. nfs4_write). Fix those cases so that the value that is eventually stored in the error field is a positive NFS status or errno, or zero. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-06-19 22:33:53 +08:00
-__entry->error,
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
show_pnfs_iomode(__entry->iomode),
(unsigned long long)__entry->offset,
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
(unsigned long long)__entry->count,
__entry->stateid_seq, __entry->stateid_hash,
__entry->layoutstateid_seq, __entry->layoutstateid_hash
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
)
);
Adding stateid information to tracepoints Operations to which stateid information is added: close, delegreturn, open, read, setattr, layoutget, layoutcommit, test_stateid, write, lock, locku, lockt Format is "stateid=<seqid>:<crc32 hash stateid.other>", also "openstateid=", "layoutstateid=", and "lockstateid=" for open_file, layoutget, set_lock tracepoints. New function is added to internal.h, nfs_stateid_hash(), to compute the hash trace_nfs4_setattr() is moved from nfs4_do_setattr() to _nfs4_do_setattr() to get access to stateid. trace_nfs4_setattr and trace_nfs4_delegreturn are changed from INODE_EVENT to new event type, INODE_STATEID_EVENT which is same as INODE_EVENT but adds stateid information for locking tracepoints, moved trace_nfs4_set_lock() into _nfs4_do_setlk() to get access to stateid information, and removed trace_nfs4_lock_reclaim(), trace_nfs4_lock_expired() as they call into _nfs4_do_setlk() and both were previously same LOCK_EVENT type. Signed-off-by: Olga Kornievskaia <kolga@netapp.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-11-25 02:29:41 +08:00
DEFINE_NFS4_INODE_STATEID_EVENT(nfs4_layoutcommit);
DEFINE_NFS4_INODE_STATEID_EVENT(nfs4_layoutreturn);
NFSv4.1/pnfs: Add a tracepoint for return-on-close events Allow tracing of return-on-close. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-08-21 04:40:47 +08:00
DEFINE_NFS4_INODE_EVENT(nfs4_layoutreturn_on_close);
NFSv4.1: Add tracepoints for debugging layoutget/return/commit Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 04:36:51 +08:00
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_UNKNOWN);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_NO_PNFS);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_RD_ZEROLEN);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_MDSTHRESH);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_NOMEM);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_BULK_RECALL);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_IO_TEST_FAIL);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_FOUND_CACHED);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_RETURN);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_BLOCKED);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_INVALID_OPEN);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_RETRY);
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET);
pnfs/flexfiles: Add tracepoints for detecting pnfs fallback to MDS Add tracepoints to allow debugging of the event chain leading to a pnfs fallback to doing I/O through the MDS. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2019-07-18 21:32:17 +08:00
TRACE_DEFINE_ENUM(PNFS_UPDATE_LAYOUT_EXIT);
NFS: Fix NFSv4 symbolic trace point output These symbolic values were not being displayed in string form. TRACE_DEFINE_ENUM was missing in many cases. It also turns out that __print_symbolic wants an unsigned long in the first field... Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2018-12-20 00:00:11 +08:00
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
#define show_pnfs_update_layout_reason(reason) \
__print_symbolic(reason, \
{ PNFS_UPDATE_LAYOUT_UNKNOWN, "unknown" }, \
{ PNFS_UPDATE_LAYOUT_NO_PNFS, "no pnfs" }, \
{ PNFS_UPDATE_LAYOUT_RD_ZEROLEN, "read+zerolen" }, \
{ PNFS_UPDATE_LAYOUT_MDSTHRESH, "mdsthresh" }, \
{ PNFS_UPDATE_LAYOUT_NOMEM, "nomem" }, \
{ PNFS_UPDATE_LAYOUT_BULK_RECALL, "bulk recall" }, \
{ PNFS_UPDATE_LAYOUT_IO_TEST_FAIL, "io test fail" }, \
{ PNFS_UPDATE_LAYOUT_FOUND_CACHED, "found cached" }, \
{ PNFS_UPDATE_LAYOUT_RETURN, "layoutreturn" }, \
{ PNFS_UPDATE_LAYOUT_BLOCKED, "layouts blocked" }, \
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
{ PNFS_UPDATE_LAYOUT_INVALID_OPEN, "invalid open" }, \
{ PNFS_UPDATE_LAYOUT_RETRY, "retrying" }, \
pnfs/flexfiles: Add tracepoints for detecting pnfs fallback to MDS Add tracepoints to allow debugging of the event chain leading to a pnfs fallback to doing I/O through the MDS. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2019-07-18 21:32:17 +08:00
{ PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET, "sent layoutget" }, \
{ PNFS_UPDATE_LAYOUT_EXIT, "exit" })
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
TRACE_EVENT(pnfs_update_layout,
TP_PROTO(struct inode *inode,
loff_t pos,
u64 count,
enum pnfs_iomode iomode,
pNFS: Modify pnfs_update_layout tracepoints to use layout stateid Instead of displaying a layout segment pointer in these tracepoints, let's use the layout stateid, now that Olga gave us a set of tools for displaying them. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-27 07:03:07 +08:00
struct pnfs_layout_hdr *lo,
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
struct pnfs_layout_segment *lseg,
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
enum pnfs_update_layout_reason reason
),
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
TP_ARGS(inode, pos, count, iomode, lo, lseg, reason),
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, fileid)
__field(u32, fhandle)
__field(loff_t, pos)
__field(u64, count)
__field(enum pnfs_iomode, iomode)
pNFS: Modify pnfs_update_layout tracepoints to use layout stateid Instead of displaying a layout segment pointer in these tracepoints, let's use the layout stateid, now that Olga gave us a set of tools for displaying them. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-27 07:03:07 +08:00
__field(int, layoutstateid_seq)
__field(u32, layoutstateid_hash)
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
__field(long, lseg)
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
__field(enum pnfs_update_layout_reason, reason)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->pos = pos;
__entry->count = count;
__entry->iomode = iomode;
__entry->reason = reason;
pNFS: Modify pnfs_update_layout tracepoints to use layout stateid Instead of displaying a layout segment pointer in these tracepoints, let's use the layout stateid, now that Olga gave us a set of tools for displaying them. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-27 07:03:07 +08:00
if (lo != NULL) {
__entry->layoutstateid_seq =
be32_to_cpu(lo->plh_stateid.seqid);
__entry->layoutstateid_hash =
nfs_stateid_hash(&lo->plh_stateid);
} else {
__entry->layoutstateid_seq = 0;
__entry->layoutstateid_hash = 0;
}
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
__entry->lseg = (long)lseg;
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
pNFS: Modify pnfs_update_layout tracepoints to use layout stateid Instead of displaying a layout segment pointer in these tracepoints, let's use the layout stateid, now that Olga gave us a set of tools for displaying them. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-27 07:03:07 +08:00
"iomode=%s pos=%llu count=%llu "
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
"layoutstateid=%d:0x%08x lseg=0x%lx (%s)",
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
show_pnfs_iomode(__entry->iomode),
(unsigned long long)__entry->pos,
pNFS: Modify pnfs_update_layout tracepoints to use layout stateid Instead of displaying a layout segment pointer in these tracepoints, let's use the layout stateid, now that Olga gave us a set of tools for displaying them. Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-27 07:03:07 +08:00
(unsigned long long)__entry->count,
__entry->layoutstateid_seq, __entry->layoutstateid_hash,
pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-18 00:28:47 +08:00
__entry->lseg,
nfs: add new tracepoint for pnfs_update_layout pnfs_update_layout is really the "nexus" of layout handling. If it returns NULL then we end up going through the MDS. This patch adds some tracepoints to that function that allow us to determine the cause when we end up going through the MDS unexpectedly. Signed-off-by: Jeff Layton <jeff.layton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-12-10 23:41:58 +08:00
show_pnfs_update_layout_reason(__entry->reason)
)
);
pnfs/flexfiles: Add tracepoints for detecting pnfs fallback to MDS Add tracepoints to allow debugging of the event chain leading to a pnfs fallback to doing I/O through the MDS. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2019-07-18 21:32:17 +08:00
DECLARE_EVENT_CLASS(pnfs_layout_event,
TP_PROTO(struct inode *inode,
loff_t pos,
u64 count,
enum pnfs_iomode iomode,
struct pnfs_layout_hdr *lo,
struct pnfs_layout_segment *lseg
),
TP_ARGS(inode, pos, count, iomode, lo, lseg),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, fileid)
__field(u32, fhandle)
__field(loff_t, pos)
__field(u64, count)
__field(enum pnfs_iomode, iomode)
__field(int, layoutstateid_seq)
__field(u32, layoutstateid_hash)
__field(long, lseg)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
__entry->pos = pos;
__entry->count = count;
__entry->iomode = iomode;
if (lo != NULL) {
__entry->layoutstateid_seq =
be32_to_cpu(lo->plh_stateid.seqid);
__entry->layoutstateid_hash =
nfs_stateid_hash(&lo->plh_stateid);
} else {
__entry->layoutstateid_seq = 0;
__entry->layoutstateid_hash = 0;
}
__entry->lseg = (long)lseg;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"iomode=%s pos=%llu count=%llu "
"layoutstateid=%d:0x%08x lseg=0x%lx",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
show_pnfs_iomode(__entry->iomode),
(unsigned long long)__entry->pos,
(unsigned long long)__entry->count,
__entry->layoutstateid_seq, __entry->layoutstateid_hash,
__entry->lseg
)
);
#define DEFINE_PNFS_LAYOUT_EVENT(name) \
DEFINE_EVENT(pnfs_layout_event, name, \
TP_PROTO(struct inode *inode, \
loff_t pos, \
u64 count, \
enum pnfs_iomode iomode, \
struct pnfs_layout_hdr *lo, \
struct pnfs_layout_segment *lseg \
), \
TP_ARGS(inode, pos, count, iomode, lo, lseg))
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_pg_init_read);
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_pg_init_write);
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_pg_get_mirror_count);
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_read_done);
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_write_done);
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_read_pagelist);
DEFINE_PNFS_LAYOUT_EVENT(pnfs_mds_fallback_write_pagelist);
pNFS/flexfiles: Add tracing for layout errors Trace layout errors for pNFS/flexfiles on read/write/commit operations. Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2020-01-07 04:25:11 +08:00
DECLARE_EVENT_CLASS(nfs4_flexfiles_io_event,
TP_PROTO(
const struct nfs_pgio_header *hdr
),
TP_ARGS(hdr),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, count)
__field(int, stateid_seq)
__field(u32, stateid_hash)
__string(dstaddr, hdr->ds_clp ?
rpc_peeraddr2str(hdr->ds_clp->cl_rpcclient,
RPC_DISPLAY_ADDR) : "unknown")
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
__entry->error = hdr->res.op_status;
__entry->fhandle = nfs_fhandle_hash(hdr->args.fh);
__entry->fileid = NFS_FILEID(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->offset = hdr->args.offset;
__entry->count = hdr->args.count;
__entry->stateid_seq =
be32_to_cpu(hdr->args.stateid.seqid);
__entry->stateid_hash =
nfs_stateid_hash(&hdr->args.stateid);
__assign_str(dstaddr, hdr->ds_clp ?
rpc_peeraddr2str(hdr->ds_clp->cl_rpcclient,
RPC_DISPLAY_ADDR) : "unknown");
),
TP_printk(
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%llu count=%u stateid=%d:0x%08x dstaddr=%s",
-__entry->error,
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->offset, __entry->count,
__entry->stateid_seq, __entry->stateid_hash,
__get_str(dstaddr)
)
);
#define DEFINE_NFS4_FLEXFILES_IO_EVENT(name) \
DEFINE_EVENT(nfs4_flexfiles_io_event, name, \
TP_PROTO( \
const struct nfs_pgio_header *hdr \
), \
TP_ARGS(hdr))
DEFINE_NFS4_FLEXFILES_IO_EVENT(ff_layout_read_error);
DEFINE_NFS4_FLEXFILES_IO_EVENT(ff_layout_write_error);
TRACE_EVENT(ff_layout_commit_error,
TP_PROTO(
const struct nfs_commit_data *data
),
TP_ARGS(data),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, count)
__string(dstaddr, data->ds_clp ?
rpc_peeraddr2str(data->ds_clp->cl_rpcclient,
RPC_DISPLAY_ADDR) : "unknown")
),
TP_fast_assign(
const struct inode *inode = data->inode;
__entry->error = data->res.op_status;
__entry->fhandle = nfs_fhandle_hash(data->args.fh);
__entry->fileid = NFS_FILEID(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->offset = data->args.offset;
__entry->count = data->args.count;
__assign_str(dstaddr, data->ds_clp ?
rpc_peeraddr2str(data->ds_clp->cl_rpcclient,
RPC_DISPLAY_ADDR) : "unknown");
),
TP_printk(
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%llu count=%u dstaddr=%s",
-__entry->error,
show_nfsv4_errors(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->offset, __entry->count,
__get_str(dstaddr)
)
);
NFSv4: Add tracepoints for debugging reads and writes Set up tracepoints to track read, write and commit, as well as pNFS reads and writes and commits to the data server. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-15 03:31:28 +08:00
#endif /* CONFIG_NFS_V4_1 */
NFSv4: Add tracepoints for debugging state management problems Set up basic tracepoints for debugging client id creation/destruction and session creation/destruction. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2013-08-09 23:51:26 +08:00
#endif /* _TRACE_NFS4_H */
#undef TRACE_INCLUDE_PATH
#define TRACE_INCLUDE_PATH .
#define TRACE_INCLUDE_FILE nfs4trace
/* This part must be outside protection */
#include <trace/define_trace.h>
Copy Permalink