linux/fs/nfs/nfstrace.h

1800 lines
45 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 14:07:57 +00:00
/* SPDX-License-Identifier: GPL-2.0 */
/*
* Copyright (c) 2013 Trond Myklebust <Trond.Myklebust@netapp.com>
*/
#undef TRACE_SYSTEM
#define TRACE_SYSTEM nfs
#if !defined(_TRACE_NFS_H) || defined(TRACE_HEADER_MULTI_READ)
#define _TRACE_NFS_H
#include <linux/tracepoint.h>
#include <linux/iversion.h>
#include <trace/misc/fs.h>
#include <trace/misc/nfs.h>
#include <trace/misc/sunrpc.h>
#define nfs_show_cache_validity(v) \
__print_flags(v, "|", \
{ NFS_INO_INVALID_DATA, "INVALID_DATA" }, \
{ NFS_INO_INVALID_ATIME, "INVALID_ATIME" }, \
{ NFS_INO_INVALID_ACCESS, "INVALID_ACCESS" }, \
{ NFS_INO_INVALID_ACL, "INVALID_ACL" }, \
{ NFS_INO_REVAL_FORCED, "REVAL_FORCED" }, \
{ NFS_INO_INVALID_LABEL, "INVALID_LABEL" }, \
{ NFS_INO_INVALID_CHANGE, "INVALID_CHANGE" }, \
{ NFS_INO_INVALID_CTIME, "INVALID_CTIME" }, \
{ NFS_INO_INVALID_MTIME, "INVALID_MTIME" }, \
{ NFS_INO_INVALID_SIZE, "INVALID_SIZE" }, \
{ NFS_INO_INVALID_OTHER, "INVALID_OTHER" }, \
{ NFS_INO_DATA_INVAL_DEFER, "DATA_INVAL_DEFER" }, \
{ NFS_INO_INVALID_BLOCKS, "INVALID_BLOCKS" }, \
{ NFS_INO_INVALID_XATTR, "INVALID_XATTR" }, \
{ NFS_INO_INVALID_NLINK, "INVALID_NLINK" }, \
{ NFS_INO_INVALID_MODE, "INVALID_MODE" })
#define nfs_show_nfsi_flags(v) \
__print_flags(v, "|", \
{ BIT(NFS_INO_STALE), "STALE" }, \
{ BIT(NFS_INO_ACL_LRU_SET), "ACL_LRU_SET" }, \
{ BIT(NFS_INO_INVALIDATING), "INVALIDATING" }, \
{ BIT(NFS_INO_FSCACHE), "FSCACHE" }, \
{ BIT(NFS_INO_LAYOUTCOMMIT), "NEED_LAYOUTCOMMIT" }, \
{ BIT(NFS_INO_LAYOUTCOMMITTING), "LAYOUTCOMMIT" }, \
{ BIT(NFS_INO_LAYOUTSTATS), "LAYOUTSTATS" }, \
{ BIT(NFS_INO_ODIRECT), "ODIRECT" })
DECLARE_EVENT_CLASS(nfs_inode_event,
TP_PROTO(
const struct inode *inode
),
TP_ARGS(inode),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, version)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->version = inode_peek_iversion_raw(inode);
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu ",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(unsigned long long)__entry->version
)
);
DECLARE_EVENT_CLASS(nfs_inode_event_done,
TP_PROTO(
const struct inode *inode,
int error
),
TP_ARGS(inode, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(dev_t, dev)
__field(u32, fhandle)
__field(unsigned char, type)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, size)
__field(unsigned long, nfsi_flags)
__field(unsigned long, cache_validity)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->error = error < 0 ? -error : 0;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->type = nfs_umode_to_dtype(inode->i_mode);
__entry->version = inode_peek_iversion_raw(inode);
__entry->size = i_size_read(inode);
__entry->nfsi_flags = nfsi->flags;
__entry->cache_validity = nfsi->cache_validity;
),
TP_printk(
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
"type=%u (%s) version=%llu size=%lld "
"cache_validity=0x%lx (%s) nfs_flags=0x%lx (%s)",
-__entry->error, show_nfs_status(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->type,
show_fs_dirent_type(__entry->type),
(unsigned long long)__entry->version,
(long long)__entry->size,
__entry->cache_validity,
nfs_show_cache_validity(__entry->cache_validity),
__entry->nfsi_flags,
nfs_show_nfsi_flags(__entry->nfsi_flags)
)
);
#define DEFINE_NFS_INODE_EVENT(name) \
DEFINE_EVENT(nfs_inode_event, name, \
TP_PROTO( \
const struct inode *inode \
), \
TP_ARGS(inode))
#define DEFINE_NFS_INODE_EVENT_DONE(name) \
DEFINE_EVENT(nfs_inode_event_done, name, \
TP_PROTO( \
const struct inode *inode, \
int error \
), \
TP_ARGS(inode, error))
DEFINE_NFS_INODE_EVENT(nfs_set_inode_stale);
DEFINE_NFS_INODE_EVENT(nfs_refresh_inode_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_refresh_inode_exit);
DEFINE_NFS_INODE_EVENT(nfs_revalidate_inode_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_revalidate_inode_exit);
DEFINE_NFS_INODE_EVENT(nfs_invalidate_mapping_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_invalidate_mapping_exit);
DEFINE_NFS_INODE_EVENT(nfs_getattr_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_getattr_exit);
DEFINE_NFS_INODE_EVENT(nfs_setattr_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_setattr_exit);
DEFINE_NFS_INODE_EVENT(nfs_writeback_page_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_writeback_page_exit);
DEFINE_NFS_INODE_EVENT(nfs_writeback_inode_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_writeback_inode_exit);
DEFINE_NFS_INODE_EVENT(nfs_fsync_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_fsync_exit);
DEFINE_NFS_INODE_EVENT(nfs_access_enter);
DEFINE_NFS_INODE_EVENT_DONE(nfs_set_cache_invalid);
DEFINE_NFS_INODE_EVENT(nfs_readdir_force_readdirplus);
DEFINE_NFS_INODE_EVENT_DONE(nfs_readdir_cache_fill_done);
DEFINE_NFS_INODE_EVENT_DONE(nfs_readdir_uncached_done);
TRACE_EVENT(nfs_access_exit,
TP_PROTO(
const struct inode *inode,
unsigned int mask,
unsigned int permitted,
int error
),
TP_ARGS(inode, mask, permitted, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(dev_t, dev)
__field(u32, fhandle)
__field(unsigned char, type)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, size)
__field(unsigned long, nfsi_flags)
__field(unsigned long, cache_validity)
__field(unsigned int, mask)
__field(unsigned int, permitted)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->error = error < 0 ? -error : 0;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->type = nfs_umode_to_dtype(inode->i_mode);
__entry->version = inode_peek_iversion_raw(inode);
__entry->size = i_size_read(inode);
__entry->nfsi_flags = nfsi->flags;
__entry->cache_validity = nfsi->cache_validity;
__entry->mask = mask;
__entry->permitted = permitted;
),
TP_printk(
"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
"type=%u (%s) version=%llu size=%lld "
"cache_validity=0x%lx (%s) nfs_flags=0x%lx (%s) "
"mask=0x%x permitted=0x%x",
-__entry->error, show_nfs_status(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
__entry->type,
show_fs_dirent_type(__entry->type),
(unsigned long long)__entry->version,
(long long)__entry->size,
__entry->cache_validity,
nfs_show_cache_validity(__entry->cache_validity),
__entry->nfsi_flags,
nfs_show_nfsi_flags(__entry->nfsi_flags),
__entry->mask, __entry->permitted
)
);
DECLARE_EVENT_CLASS(nfs_update_size_class,
TP_PROTO(
const struct inode *inode,
loff_t new_size
),
TP_ARGS(inode, new_size),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, cur_size)
__field(loff_t, new_size)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->fileid = nfsi->fileid;
__entry->version = inode_peek_iversion_raw(inode);
__entry->cur_size = i_size_read(inode);
__entry->new_size = new_size;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu cursize=%lld newsize=%lld",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->version,
__entry->cur_size, __entry->new_size
)
);
#define DEFINE_NFS_UPDATE_SIZE_EVENT(name) \
DEFINE_EVENT(nfs_update_size_class, nfs_size_##name, \
TP_PROTO( \
const struct inode *inode, \
loff_t new_size \
), \
TP_ARGS(inode, new_size))
DEFINE_NFS_UPDATE_SIZE_EVENT(truncate);
DEFINE_NFS_UPDATE_SIZE_EVENT(wcc);
DEFINE_NFS_UPDATE_SIZE_EVENT(update);
DEFINE_NFS_UPDATE_SIZE_EVENT(grow);
DECLARE_EVENT_CLASS(nfs_inode_range_event,
TP_PROTO(
const struct inode *inode,
loff_t range_start,
loff_t range_end
),
TP_ARGS(inode, range_start, range_end),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, range_start)
__field(loff_t, range_end)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->fileid = nfsi->fileid;
__entry->version = inode_peek_iversion_raw(inode);
__entry->range_start = range_start;
__entry->range_end = range_end;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu "
"range=[%lld, %lld]",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->version,
__entry->range_start, __entry->range_end
)
);
#define DEFINE_NFS_INODE_RANGE_EVENT(name) \
DEFINE_EVENT(nfs_inode_range_event, name, \
TP_PROTO( \
const struct inode *inode, \
loff_t range_start, \
loff_t range_end \
), \
TP_ARGS(inode, range_start, range_end))
DEFINE_NFS_INODE_RANGE_EVENT(nfs_readdir_invalidate_cache_range);
DECLARE_EVENT_CLASS(nfs_readdir_event,
TP_PROTO(
const struct file *file,
const __be32 *verifier,
u64 cookie,
pgoff_t page_index,
unsigned int dtsize
),
TP_ARGS(file, verifier, cookie, page_index, dtsize),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, version)
__array(char, verifier, NFS4_VERIFIER_SIZE)
__field(u64, cookie)
__field(pgoff_t, index)
__field(unsigned int, dtsize)
),
TP_fast_assign(
const struct inode *dir = file_inode(file);
const struct nfs_inode *nfsi = NFS_I(dir);
__entry->dev = dir->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->version = inode_peek_iversion_raw(dir);
if (cookie != 0)
memcpy(__entry->verifier, verifier,
NFS4_VERIFIER_SIZE);
else
memset(__entry->verifier, 0,
NFS4_VERIFIER_SIZE);
__entry->cookie = cookie;
__entry->index = page_index;
__entry->dtsize = dtsize;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu "
"cookie=%s:0x%llx cache_index=%lu dtsize=%u",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid, __entry->fhandle,
__entry->version, show_nfs4_verifier(__entry->verifier),
(unsigned long long)__entry->cookie, __entry->index,
__entry->dtsize
)
);
#define DEFINE_NFS_READDIR_EVENT(name) \
DEFINE_EVENT(nfs_readdir_event, name, \
TP_PROTO( \
const struct file *file, \
const __be32 *verifier, \
u64 cookie, \
pgoff_t page_index, \
unsigned int dtsize \
), \
TP_ARGS(file, verifier, cookie, page_index, dtsize))
DEFINE_NFS_READDIR_EVENT(nfs_readdir_cache_fill);
DEFINE_NFS_READDIR_EVENT(nfs_readdir_uncached);
DECLARE_EVENT_CLASS(nfs_lookup_event,
TP_PROTO(
const struct inode *dir,
const struct dentry *dentry,
unsigned int flags
),
TP_ARGS(dir, dentry, flags),
TP_STRUCT__entry(
__field(unsigned long, flags)
__field(dev_t, dev)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->flags = flags;
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"flags=0x%lx (%s) name=%02x:%02x:%llu/%s",
__entry->flags,
show_fs_lookup_flags(__entry->flags),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
#define DEFINE_NFS_LOOKUP_EVENT(name) \
DEFINE_EVENT(nfs_lookup_event, name, \
TP_PROTO( \
const struct inode *dir, \
const struct dentry *dentry, \
unsigned int flags \
), \
TP_ARGS(dir, dentry, flags))
DECLARE_EVENT_CLASS(nfs_lookup_event_done,
TP_PROTO(
const struct inode *dir,
const struct dentry *dentry,
unsigned int flags,
int error
),
TP_ARGS(dir, dentry, flags, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(unsigned long, flags)
__field(dev_t, dev)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->error = error < 0 ? -error : 0;
__entry->flags = flags;
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"error=%ld (%s) flags=0x%lx (%s) name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
__entry->flags,
show_fs_lookup_flags(__entry->flags),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
#define DEFINE_NFS_LOOKUP_EVENT_DONE(name) \
DEFINE_EVENT(nfs_lookup_event_done, name, \
TP_PROTO( \
const struct inode *dir, \
const struct dentry *dentry, \
unsigned int flags, \
int error \
), \
TP_ARGS(dir, dentry, flags, error))
DEFINE_NFS_LOOKUP_EVENT(nfs_lookup_enter);
DEFINE_NFS_LOOKUP_EVENT_DONE(nfs_lookup_exit);
DEFINE_NFS_LOOKUP_EVENT(nfs_lookup_revalidate_enter);
DEFINE_NFS_LOOKUP_EVENT_DONE(nfs_lookup_revalidate_exit);
DEFINE_NFS_LOOKUP_EVENT(nfs_readdir_lookup);
DEFINE_NFS_LOOKUP_EVENT(nfs_readdir_lookup_revalidate_failed);
DEFINE_NFS_LOOKUP_EVENT_DONE(nfs_readdir_lookup_revalidate);
TRACE_EVENT(nfs_atomic_open_enter,
TP_PROTO(
const struct inode *dir,
const struct nfs_open_context *ctx,
unsigned int flags
),
TP_ARGS(dir, ctx, flags),
TP_STRUCT__entry(
__field(unsigned long, flags)
__field(unsigned long, fmode)
__field(dev_t, dev)
__field(u64, dir)
__string(name, ctx->dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->flags = flags;
__entry->fmode = (__force unsigned long)ctx->mode;
__assign_str(name, ctx->dentry->d_name.name);
),
TP_printk(
"flags=0x%lx (%s) fmode=%s name=%02x:%02x:%llu/%s",
__entry->flags,
show_fs_fcntl_open_flags(__entry->flags),
show_fs_fmode_flags(__entry->fmode),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
TRACE_EVENT(nfs_atomic_open_exit,
TP_PROTO(
const struct inode *dir,
const struct nfs_open_context *ctx,
unsigned int flags,
int error
),
TP_ARGS(dir, ctx, flags, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(unsigned long, flags)
__field(unsigned long, fmode)
__field(dev_t, dev)
__field(u64, dir)
__string(name, ctx->dentry->d_name.name)
),
TP_fast_assign(
__entry->error = -error;
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->flags = flags;
__entry->fmode = (__force unsigned long)ctx->mode;
__assign_str(name, ctx->dentry->d_name.name);
),
TP_printk(
"error=%ld (%s) flags=0x%lx (%s) fmode=%s "
"name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
__entry->flags,
show_fs_fcntl_open_flags(__entry->flags),
show_fs_fmode_flags(__entry->fmode),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
TRACE_EVENT(nfs_create_enter,
TP_PROTO(
const struct inode *dir,
const struct dentry *dentry,
unsigned int flags
),
TP_ARGS(dir, dentry, flags),
TP_STRUCT__entry(
__field(unsigned long, flags)
__field(dev_t, dev)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->flags = flags;
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"flags=0x%lx (%s) name=%02x:%02x:%llu/%s",
__entry->flags,
show_fs_fcntl_open_flags(__entry->flags),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
TRACE_EVENT(nfs_create_exit,
TP_PROTO(
const struct inode *dir,
const struct dentry *dentry,
unsigned int flags,
int error
),
TP_ARGS(dir, dentry, flags, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(unsigned long, flags)
__field(dev_t, dev)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->error = -error;
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->flags = flags;
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"error=%ld (%s) flags=0x%lx (%s) name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
__entry->flags,
show_fs_fcntl_open_flags(__entry->flags),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
DECLARE_EVENT_CLASS(nfs_directory_event,
TP_PROTO(
const struct inode *dir,
const struct dentry *dentry
),
TP_ARGS(dir, dentry),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"name=%02x:%02x:%llu/%s",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
#define DEFINE_NFS_DIRECTORY_EVENT(name) \
DEFINE_EVENT(nfs_directory_event, name, \
TP_PROTO( \
const struct inode *dir, \
const struct dentry *dentry \
), \
TP_ARGS(dir, dentry))
DECLARE_EVENT_CLASS(nfs_directory_event_done,
TP_PROTO(
const struct inode *dir,
const struct dentry *dentry,
int error
),
TP_ARGS(dir, dentry, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(dev_t, dev)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->error = error < 0 ? -error : 0;
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"error=%ld (%s) name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
#define DEFINE_NFS_DIRECTORY_EVENT_DONE(name) \
DEFINE_EVENT(nfs_directory_event_done, name, \
TP_PROTO( \
const struct inode *dir, \
const struct dentry *dentry, \
int error \
), \
TP_ARGS(dir, dentry, error))
DEFINE_NFS_DIRECTORY_EVENT(nfs_mknod_enter);
DEFINE_NFS_DIRECTORY_EVENT_DONE(nfs_mknod_exit);
DEFINE_NFS_DIRECTORY_EVENT(nfs_mkdir_enter);
DEFINE_NFS_DIRECTORY_EVENT_DONE(nfs_mkdir_exit);
DEFINE_NFS_DIRECTORY_EVENT(nfs_rmdir_enter);
DEFINE_NFS_DIRECTORY_EVENT_DONE(nfs_rmdir_exit);
DEFINE_NFS_DIRECTORY_EVENT(nfs_remove_enter);
DEFINE_NFS_DIRECTORY_EVENT_DONE(nfs_remove_exit);
DEFINE_NFS_DIRECTORY_EVENT(nfs_unlink_enter);
DEFINE_NFS_DIRECTORY_EVENT_DONE(nfs_unlink_exit);
DEFINE_NFS_DIRECTORY_EVENT(nfs_symlink_enter);
DEFINE_NFS_DIRECTORY_EVENT_DONE(nfs_symlink_exit);
TRACE_EVENT(nfs_link_enter,
TP_PROTO(
const struct inode *inode,
const struct inode *dir,
const struct dentry *dentry
),
TP_ARGS(inode, dir, dentry),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, fileid)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->dir = NFS_FILEID(dir);
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"fileid=%02x:%02x:%llu name=%02x:%02x:%llu/%s",
MAJOR(__entry->dev), MINOR(__entry->dev),
__entry->fileid,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
TRACE_EVENT(nfs_link_exit,
TP_PROTO(
const struct inode *inode,
const struct inode *dir,
const struct dentry *dentry,
int error
),
TP_ARGS(inode, dir, dentry, error),
TP_STRUCT__entry(
__field(unsigned long, error)
__field(dev_t, dev)
__field(u64, fileid)
__field(u64, dir)
__string(name, dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = NFS_FILEID(inode);
__entry->dir = NFS_FILEID(dir);
__entry->error = error < 0 ? -error : 0;
__assign_str(name, dentry->d_name.name);
),
TP_printk(
"error=%ld (%s) fileid=%02x:%02x:%llu name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
__entry->fileid,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
DECLARE_EVENT_CLASS(nfs_rename_event,
TP_PROTO(
const struct inode *old_dir,
const struct dentry *old_dentry,
const struct inode *new_dir,
const struct dentry *new_dentry
),
TP_ARGS(old_dir, old_dentry, new_dir, new_dentry),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, old_dir)
__field(u64, new_dir)
__string(old_name, old_dentry->d_name.name)
__string(new_name, new_dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = old_dir->i_sb->s_dev;
__entry->old_dir = NFS_FILEID(old_dir);
__entry->new_dir = NFS_FILEID(new_dir);
__assign_str(old_name, old_dentry->d_name.name);
__assign_str(new_name, new_dentry->d_name.name);
),
TP_printk(
"old_name=%02x:%02x:%llu/%s new_name=%02x:%02x:%llu/%s",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->old_dir,
__get_str(old_name),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->new_dir,
__get_str(new_name)
)
);
#define DEFINE_NFS_RENAME_EVENT(name) \
DEFINE_EVENT(nfs_rename_event, name, \
TP_PROTO( \
const struct inode *old_dir, \
const struct dentry *old_dentry, \
const struct inode *new_dir, \
const struct dentry *new_dentry \
), \
TP_ARGS(old_dir, old_dentry, new_dir, new_dentry))
DECLARE_EVENT_CLASS(nfs_rename_event_done,
TP_PROTO(
const struct inode *old_dir,
const struct dentry *old_dentry,
const struct inode *new_dir,
const struct dentry *new_dentry,
int error
),
TP_ARGS(old_dir, old_dentry, new_dir, new_dentry, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(unsigned long, error)
__field(u64, old_dir)
__string(old_name, old_dentry->d_name.name)
__field(u64, new_dir)
__string(new_name, new_dentry->d_name.name)
),
TP_fast_assign(
__entry->dev = old_dir->i_sb->s_dev;
__entry->error = -error;
__entry->old_dir = NFS_FILEID(old_dir);
__entry->new_dir = NFS_FILEID(new_dir);
__assign_str(old_name, old_dentry->d_name.name);
__assign_str(new_name, new_dentry->d_name.name);
),
TP_printk(
"error=%ld (%s) old_name=%02x:%02x:%llu/%s "
"new_name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->old_dir,
__get_str(old_name),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->new_dir,
__get_str(new_name)
)
);
#define DEFINE_NFS_RENAME_EVENT_DONE(name) \
DEFINE_EVENT(nfs_rename_event_done, name, \
TP_PROTO( \
const struct inode *old_dir, \
const struct dentry *old_dentry, \
const struct inode *new_dir, \
const struct dentry *new_dentry, \
int error \
), \
TP_ARGS(old_dir, old_dentry, new_dir, \
new_dentry, error))
DEFINE_NFS_RENAME_EVENT(nfs_rename_enter);
DEFINE_NFS_RENAME_EVENT_DONE(nfs_rename_exit);
DEFINE_NFS_RENAME_EVENT_DONE(nfs_sillyrename_rename);
TRACE_EVENT(nfs_sillyrename_unlink,
TP_PROTO(
const struct nfs_unlinkdata *data,
int error
),
TP_ARGS(data, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(unsigned long, error)
__field(u64, dir)
__dynamic_array(char, name, data->args.name.len + 1)
),
TP_fast_assign(
struct inode *dir = d_inode(data->dentry->d_parent);
size_t len = data->args.name.len;
__entry->dev = dir->i_sb->s_dev;
__entry->dir = NFS_FILEID(dir);
__entry->error = -error;
memcpy(__get_str(name),
data->args.name.name, len);
__get_str(name)[len] = 0;
),
TP_printk(
"error=%ld (%s) name=%02x:%02x:%llu/%s",
-__entry->error, show_nfs_status(__entry->error),
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->dir,
__get_str(name)
)
);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TRACE_EVENT(nfs_aop_readpage,
TP_PROTO(
const struct inode *inode,
struct page *page
),
TP_ARGS(inode, page),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, offset)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->version = inode_peek_iversion_raw(inode);
__entry->offset = page_index(page) << PAGE_SHIFT;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu offset=%lld",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->version,
__entry->offset
)
);
TRACE_EVENT(nfs_aop_readpage_done,
TP_PROTO(
const struct inode *inode,
struct page *page,
int ret
),
TP_ARGS(inode, page, ret),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(int, ret)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, offset)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->version = inode_peek_iversion_raw(inode);
__entry->offset = page_index(page) << PAGE_SHIFT;
__entry->ret = ret;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu offset=%lld ret=%d",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->version,
__entry->offset, __entry->ret
)
);
TRACE_EVENT(nfs_aop_readahead,
TP_PROTO(
const struct inode *inode,
loff_t pos,
unsigned int nr_pages
),
TP_ARGS(inode, pos, nr_pages),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, offset)
__field(unsigned int, nr_pages)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->version = inode_peek_iversion_raw(inode);
__entry->offset = pos;
__entry->nr_pages = nr_pages;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu offset=%lld nr_pages=%u",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->version,
__entry->offset, __entry->nr_pages
)
);
TRACE_EVENT(nfs_aop_readahead_done,
TP_PROTO(
const struct inode *inode,
unsigned int nr_pages,
int ret
),
TP_ARGS(inode, nr_pages, ret),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(int, ret)
__field(u64, fileid)
__field(u64, version)
__field(loff_t, offset)
__field(unsigned int, nr_pages)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->version = inode_peek_iversion_raw(inode);
__entry->nr_pages = nr_pages;
__entry->ret = ret;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x version=%llu nr_pages=%u ret=%d",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->version,
__entry->nr_pages, __entry->ret
)
);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TRACE_EVENT(nfs_initiate_read,
TP_PROTO(
const struct nfs_pgio_header *hdr
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_ARGS(hdr),
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, count)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->offset = hdr->args.offset;
__entry->count = hdr->args.count;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u",
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->count
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
)
);
TRACE_EVENT(nfs_readpage_done,
TP_PROTO(
const struct rpc_task *task,
const struct nfs_pgio_header *hdr
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_ARGS(task, hdr),
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, arg_count)
__field(u32, res_count)
__field(bool, eof)
__field(int, error)
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
__entry->error = task->tk_status;
__entry->offset = hdr->args.offset;
__entry->arg_count = hdr->args.count;
__entry->res_count = hdr->res.count;
__entry->eof = hdr->res.eof;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
),
TP_printk(
"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u res=%u%s", __entry->error,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->arg_count,
__entry->res_count, __entry->eof ? " eof" : ""
)
);
TRACE_EVENT(nfs_readpage_short,
TP_PROTO(
const struct rpc_task *task,
const struct nfs_pgio_header *hdr
),
TP_ARGS(task, hdr),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, arg_count)
__field(u32, res_count)
__field(bool, eof)
__field(int, error)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
__entry->error = task->tk_status;
__entry->offset = hdr->args.offset;
__entry->arg_count = hdr->args.count;
__entry->res_count = hdr->res.count;
__entry->eof = hdr->res.eof;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_printk(
"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u res=%u%s", __entry->error,
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->arg_count,
__entry->res_count, __entry->eof ? " eof" : ""
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
)
);
DECLARE_EVENT_CLASS(nfs_fscache_page_event,
TP_PROTO(
const struct inode *inode,
struct page *page
),
TP_ARGS(inode, page),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = &nfsi->fh;
__entry->offset = page_index(page) << PAGE_SHIFT;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset
)
);
DECLARE_EVENT_CLASS(nfs_fscache_page_event_done,
TP_PROTO(
const struct inode *inode,
struct page *page,
int error
),
TP_ARGS(inode, page, error),
TP_STRUCT__entry(
__field(int, error)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = &nfsi->fh;
__entry->offset = page_index(page) << PAGE_SHIFT;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
__entry->error = error;
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld error=%d",
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->error
)
);
#define DEFINE_NFS_FSCACHE_PAGE_EVENT(name) \
DEFINE_EVENT(nfs_fscache_page_event, name, \
TP_PROTO( \
const struct inode *inode, \
struct page *page \
), \
TP_ARGS(inode, page))
#define DEFINE_NFS_FSCACHE_PAGE_EVENT_DONE(name) \
DEFINE_EVENT(nfs_fscache_page_event_done, name, \
TP_PROTO( \
const struct inode *inode, \
struct page *page, \
int error \
), \
TP_ARGS(inode, page, error))
DEFINE_NFS_FSCACHE_PAGE_EVENT(nfs_fscache_read_page);
DEFINE_NFS_FSCACHE_PAGE_EVENT_DONE(nfs_fscache_read_page_exit);
DEFINE_NFS_FSCACHE_PAGE_EVENT(nfs_fscache_write_page);
DEFINE_NFS_FSCACHE_PAGE_EVENT_DONE(nfs_fscache_write_page_exit);
TRACE_EVENT(nfs_pgio_error,
TP_PROTO(
const struct nfs_pgio_header *hdr,
int error,
loff_t pos
),
TP_ARGS(hdr, error, pos),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, arg_count)
__field(u32, res_count)
__field(loff_t, pos)
__field(int, error)
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
__entry->error = error;
__entry->offset = hdr->args.offset;
__entry->arg_count = hdr->args.count;
__entry->res_count = hdr->res.count;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
),
TP_printk("error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u res=%u pos=%llu", __entry->error,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid, __entry->fhandle,
(long long)__entry->offset, __entry->arg_count, __entry->res_count,
__entry->pos
)
);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TRACE_EVENT(nfs_initiate_write,
TP_PROTO(
const struct nfs_pgio_header *hdr
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_ARGS(hdr),
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, count)
__field(unsigned long, stable)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->offset = hdr->args.offset;
__entry->count = hdr->args.count;
__entry->stable = hdr->args.stable;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u stable=%s",
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->count,
show_nfs_stable_how(__entry->stable)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
)
);
TRACE_EVENT(nfs_writeback_done,
TP_PROTO(
const struct rpc_task *task,
const struct nfs_pgio_header *hdr
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_ARGS(task, hdr),
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, arg_count)
__field(u32, res_count)
__field(int, error)
__field(unsigned long, stable)
__array(char, verifier, NFS4_VERIFIER_SIZE)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_fast_assign(
const struct inode *inode = hdr->inode;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = hdr->args.fh ?
hdr->args.fh : &nfsi->fh;
const struct nfs_writeverf *verf = hdr->res.verf;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->error = task->tk_status;
__entry->offset = hdr->args.offset;
__entry->arg_count = hdr->args.count;
__entry->res_count = hdr->res.count;
__entry->stable = verf->committed;
memcpy(__entry->verifier,
&verf->verifier,
NFS4_VERIFIER_SIZE);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_printk(
"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u res=%u stable=%s "
"verifier=%s", __entry->error,
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->arg_count,
__entry->res_count,
show_nfs_stable_how(__entry->stable),
show_nfs4_verifier(__entry->verifier)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
)
);
DECLARE_EVENT_CLASS(nfs_page_error_class,
TP_PROTO(
const struct inode *inode,
const struct nfs_page *req,
int error
),
TP_ARGS(inode, req, error),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(unsigned int, count)
__field(int, error)
),
TP_fast_assign(
const struct nfs_inode *nfsi = NFS_I(inode);
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(&nfsi->fh);
__entry->offset = req_offset(req);
__entry->count = req->wb_bytes;
__entry->error = error;
),
TP_printk(
"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u", __entry->error,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->offset,
__entry->count
)
);
#define DEFINE_NFS_PAGEERR_EVENT(name) \
DEFINE_EVENT(nfs_page_error_class, name, \
TP_PROTO( \
const struct inode *inode, \
const struct nfs_page *req, \
int error \
), \
TP_ARGS(inode, req, error))
DEFINE_NFS_PAGEERR_EVENT(nfs_write_error);
DEFINE_NFS_PAGEERR_EVENT(nfs_comp_error);
DEFINE_NFS_PAGEERR_EVENT(nfs_commit_error);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TRACE_EVENT(nfs_initiate_commit,
TP_PROTO(
const struct nfs_commit_data *data
),
TP_ARGS(data),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(u32, count)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_fast_assign(
const struct inode *inode = data->inode;
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = data->args.fh ?
data->args.fh : &nfsi->fh;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->offset = data->args.offset;
__entry->count = data->args.count;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_printk(
"fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%u",
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset, __entry->count
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
)
);
TRACE_EVENT(nfs_commit_done,
TP_PROTO(
const struct rpc_task *task,
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
const struct nfs_commit_data *data
),
TP_ARGS(task, data),
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
__field(loff_t, offset)
__field(int, error)
__field(unsigned long, stable)
__array(char, verifier, NFS4_VERIFIER_SIZE)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_fast_assign(
const struct inode *inode = data->inode;
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = data->args.fh ?
data->args.fh : &nfsi->fh;
const struct nfs_writeverf *verf = data->res.verf;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->error = task->tk_status;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->offset = data->args.offset;
__entry->stable = verf->committed;
memcpy(__entry->verifier,
&verf->verifier,
NFS4_VERIFIER_SIZE);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
),
TP_printk(
"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld stable=%s verifier=%s", __entry->error,
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle,
(long long)__entry->offset,
show_nfs_stable_how(__entry->stable),
show_nfs4_verifier(__entry->verifier)
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
)
);
#define nfs_show_direct_req_flags(v) \
__print_flags(v, "|", \
{ NFS_ODIRECT_DO_COMMIT, "DO_COMMIT" }, \
{ NFS_ODIRECT_RESCHED_WRITES, "RESCHED_WRITES" }, \
{ NFS_ODIRECT_SHOULD_DIRTY, "SHOULD DIRTY" }, \
{ NFS_ODIRECT_DONE, "DONE" } )
DECLARE_EVENT_CLASS(nfs_direct_req_class,
TP_PROTO(
const struct nfs_direct_req *dreq
),
TP_ARGS(dreq),
TP_STRUCT__entry(
__field(dev_t, dev)
__field(u64, fileid)
__field(u32, fhandle)
__field(loff_t, offset)
__field(ssize_t, count)
__field(ssize_t, bytes_left)
__field(ssize_t, error)
__field(int, flags)
),
TP_fast_assign(
const struct inode *inode = dreq->inode;
const struct nfs_inode *nfsi = NFS_I(inode);
const struct nfs_fh *fh = &nfsi->fh;
__entry->dev = inode->i_sb->s_dev;
__entry->fileid = nfsi->fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
__entry->offset = dreq->io_start;
__entry->count = dreq->count;
__entry->bytes_left = dreq->bytes_left;
__entry->error = dreq->error;
__entry->flags = dreq->flags;
),
TP_printk(
"error=%zd fileid=%02x:%02x:%llu fhandle=0x%08x "
"offset=%lld count=%zd bytes_left=%zd flags=%s",
__entry->error, MAJOR(__entry->dev),
MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle, __entry->offset,
__entry->count, __entry->bytes_left,
nfs_show_direct_req_flags(__entry->flags)
)
);
#define DEFINE_NFS_DIRECT_REQ_EVENT(name) \
DEFINE_EVENT(nfs_direct_req_class, name, \
TP_PROTO( \
const struct nfs_direct_req *dreq \
), \
TP_ARGS(dreq))
DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_commit_complete);
DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_resched_write);
DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_complete);
DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_completion);
DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_schedule_iovec);
DEFINE_NFS_DIRECT_REQ_EVENT(nfs_direct_write_reschedule_io);
TRACE_EVENT(nfs_fh_to_dentry,
TP_PROTO(
const struct super_block *sb,
const struct nfs_fh *fh,
u64 fileid,
int error
),
TP_ARGS(sb, fh, fileid, error),
TP_STRUCT__entry(
__field(int, error)
__field(dev_t, dev)
__field(u32, fhandle)
__field(u64, fileid)
),
TP_fast_assign(
__entry->error = error;
__entry->dev = sb->s_dev;
__entry->fileid = fileid;
__entry->fhandle = nfs_fhandle_hash(fh);
),
TP_printk(
"error=%d fileid=%02x:%02x:%llu fhandle=0x%08x ",
__entry->error,
MAJOR(__entry->dev), MINOR(__entry->dev),
(unsigned long long)__entry->fileid,
__entry->fhandle
)
);
TRACE_EVENT(nfs_mount_assign,
TP_PROTO(
const char *option,
const char *value
),
TP_ARGS(option, value),
TP_STRUCT__entry(
__string(option, option)
__string(value, value)
),
TP_fast_assign(
__assign_str(option, option);
__assign_str(value, value);
),
TP_printk("option %s=%s",
__get_str(option), __get_str(value)
)
);
TRACE_EVENT(nfs_mount_option,
TP_PROTO(
const struct fs_parameter *param
),
TP_ARGS(param),
TP_STRUCT__entry(
__string(option, param->key)
),
TP_fast_assign(
__assign_str(option, param->key);
),
TP_printk("option %s", __get_str(option))
);
TRACE_EVENT(nfs_mount_path,
TP_PROTO(
const char *path
),
TP_ARGS(path),
TP_STRUCT__entry(
__string(path, path)
),
TP_fast_assign(
__assign_str(path, path);
),
TP_printk("path='%s'", __get_str(path))
);
DECLARE_EVENT_CLASS(nfs_xdr_event,
TP_PROTO(
const struct xdr_stream *xdr,
int error
),
TP_ARGS(xdr, error),
TP_STRUCT__entry(
__field(unsigned int, task_id)
__field(unsigned int, client_id)
__field(u32, xid)
__field(int, version)
__field(unsigned long, error)
__string(program,
xdr->rqst->rq_task->tk_client->cl_program->name)
__string(procedure,
xdr->rqst->rq_task->tk_msg.rpc_proc->p_name)
),
TP_fast_assign(
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);
__entry->version = task->tk_client->cl_vers;
__entry->error = error;
__assign_str(program,
task->tk_client->cl_program->name);
__assign_str(procedure, task->tk_msg.rpc_proc->p_name);
),
TP_printk(SUNRPC_TRACE_TASK_SPECIFIER
" xid=0x%08x %sv%d %s error=%ld (%s)",
__entry->task_id, __entry->client_id, __entry->xid,
__get_str(program), __entry->version,
__get_str(procedure), -__entry->error,
show_nfs_status(__entry->error)
)
);
#define DEFINE_NFS_XDR_EVENT(name) \
DEFINE_EVENT(nfs_xdr_event, name, \
TP_PROTO( \
const struct xdr_stream *xdr, \
int error \
), \
TP_ARGS(xdr, error))
DEFINE_NFS_XDR_EVENT(nfs_xdr_status);
DEFINE_NFS_XDR_EVENT(nfs_xdr_bad_filehandle);
#endif /* _TRACE_NFS_H */
#undef TRACE_INCLUDE_PATH
#define TRACE_INCLUDE_PATH .
#define TRACE_INCLUDE_FILE nfstrace
/* This part must be outside protection */
#include <trace/define_trace.h>