openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/include/trace/events/afs.h

1365 lines
40 KiB
C
Raw Normal View History

treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 36 Based on 1 normalized pattern(s): this program is free software you can redistribute it and or modify it under the terms of the gnu general public licence as published by the free software foundation either version 2 of the licence or at your option any later version extracted by the scancode license scanner the SPDX license identifier GPL-2.0-or-later has been chosen to replace the boilerplate/reference in 114 file(s). Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Allison Randal <allison@lohutok.net> Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Cc: linux-spdx@vger.kernel.org Link: https://lkml.kernel.org/r/20190520170857.552531963@linutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2019-05-21 01:08:01 +08:00
/* SPDX-License-Identifier: GPL-2.0-or-later */
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
/* AFS tracepoints
*
* Copyright (C) 2016 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#undef TRACE_SYSTEM
#define TRACE_SYSTEM afs
#if !defined(_TRACE_AFS_H) || defined(TRACE_HEADER_MULTI_READ)
#define _TRACE_AFS_H
#include <linux/tracepoint.h>
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
/*
* Define enums for tracing information.
*/
#ifndef __AFS_DECLARE_TRACE_ENUMS_ONCE_ONLY
#define __AFS_DECLARE_TRACE_ENUMS_ONCE_ONLY
enum afs_call_trace {
afs_call_trace_alloc,
afs_call_trace_free,
afs: Fix race in async call refcounting There's a race between afs_make_call() and afs_wake_up_async_call() in the case that an error is returned from rxrpc_kernel_send_data() after it has queued the final packet. afs_make_call() will try and clean up the mess, but the call state may have been moved on thereby causing afs_process_async_call() to also try and to delete the call. Fix this by: (1) Getting an extra ref for an asynchronous call for the call itself to hold. This makes sure the call doesn't evaporate on us accidentally and will allow the call to be retained by the caller in a future patch. The ref is released on leaving afs_make_call() or afs_wait_for_call_to_complete(). (2) In the event of an error from rxrpc_kernel_send_data(): (a) Don't set the call state to AFS_CALL_COMPLETE until *after* the call has been aborted and ended. This prevents afs_deliver_to_call() from doing anything with any notifications it gets. (b) Explicitly end the call immediately to prevent further callbacks. (c) Cancel any queued async_work and wait for the work if it's executing. This allows us to be sure the race won't recur when we change the state. We put the work queue's ref on the call if we managed to cancel it. (d) Put the call's ref that we got in (1). This belongs to us as long as the call is in state AFS_CALL_CL_REQUESTING. Fixes: 341f741f04be ("afs: Refcount the afs_call struct") Signed-off-by: David Howells <dhowells@redhat.com>
2019-01-10 23:40:50 +08:00
afs_call_trace_get,
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
afs_call_trace_put,
afs_call_trace_wake,
afs_call_trace_work,
};
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
enum afs_server_trace {
afs_server_trace_alloc,
afs_server_trace_callback,
afs_server_trace_destroy,
afs_server_trace_free,
afs_server_trace_gc,
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
afs_server_trace_get_by_addr,
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
afs_server_trace_get_by_uuid,
afs_server_trace_get_caps,
afs_server_trace_get_install,
afs_server_trace_get_new_cbi,
afs: Actively poll fileservers to maintain NAT or firewall openings When an AFS client accesses a file, it receives a limited-duration callback promise that the server will notify it if another client changes a file. This callback duration can be a few hours in length. If a client mounts a volume and then an application prevents it from being unmounted, say by chdir'ing into it, but then does nothing for some time, the rxrpc_peer record will expire and rxrpc-level keepalive will cease. If there is NAT or a firewall between the client and the server, the route back for the server may close after a comparatively short duration, meaning that attempts by the server to notify the client may then bounce. The client, however, may (so far as it knows) still have a valid unexpired promise and will then rely on its cached data and will not see changes made on the server by a third party until it incidentally rechecks the status or the promise needs renewal. To deal with this, the client needs to regularly probe the server. This has two effects: firstly, it keeps a route open back for the server, and secondly, it causes the server to disgorge any notifications that got queued up because they couldn't be sent. Fix this by adding a mechanism to emit regular probes. Two levels of probing are made available: Under normal circumstances the 'slow' queue will be used for a fileserver - this just probes the preferred address once every 5 mins or so; however, if server fails to respond to any probes, the server will shift to the 'fast' queue from which all its interfaces will be probed every 30s. When it finally responds, the record will switch back to the slow queue. Further notes: (1) Probing is now no longer driven from the fileserver rotation algorithm. (2) Probes are dispatched to all interfaces on a fileserver when that an afs_server object is set up to record it. (3) The afs_server object is removed from the probe queues when we start to probe it. afs_is_probing_server() returns true if it's not listed - ie. it's undergoing probing. (4) The afs_server object is added back on to the probe queue when the final outstanding probe completes, but the probed_at time is set when we're about to launch a probe so that it's not dependent on the probe duration. (5) The timer and the work item added for this must be handed a count on net->servers_outstanding, which they hand on or release. This makes sure that network namespace cleanup waits for them. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Reported-by: Dave Botsch <botsch@cnf.cornell.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-24 22:10:00 +08:00
afs_server_trace_get_probe,
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
afs_server_trace_give_up_cb,
afs_server_trace_put_call,
afs_server_trace_put_cbi,
afs_server_trace_put_find_rsq,
afs: Actively poll fileservers to maintain NAT or firewall openings When an AFS client accesses a file, it receives a limited-duration callback promise that the server will notify it if another client changes a file. This callback duration can be a few hours in length. If a client mounts a volume and then an application prevents it from being unmounted, say by chdir'ing into it, but then does nothing for some time, the rxrpc_peer record will expire and rxrpc-level keepalive will cease. If there is NAT or a firewall between the client and the server, the route back for the server may close after a comparatively short duration, meaning that attempts by the server to notify the client may then bounce. The client, however, may (so far as it knows) still have a valid unexpired promise and will then rely on its cached data and will not see changes made on the server by a third party until it incidentally rechecks the status or the promise needs renewal. To deal with this, the client needs to regularly probe the server. This has two effects: firstly, it keeps a route open back for the server, and secondly, it causes the server to disgorge any notifications that got queued up because they couldn't be sent. Fix this by adding a mechanism to emit regular probes. Two levels of probing are made available: Under normal circumstances the 'slow' queue will be used for a fileserver - this just probes the preferred address once every 5 mins or so; however, if server fails to respond to any probes, the server will shift to the 'fast' queue from which all its interfaces will be probed every 30s. When it finally responds, the record will switch back to the slow queue. Further notes: (1) Probing is now no longer driven from the fileserver rotation algorithm. (2) Probes are dispatched to all interfaces on a fileserver when that an afs_server object is set up to record it. (3) The afs_server object is removed from the probe queues when we start to probe it. afs_is_probing_server() returns true if it's not listed - ie. it's undergoing probing. (4) The afs_server object is added back on to the probe queue when the final outstanding probe completes, but the probed_at time is set when we're about to launch a probe so that it's not dependent on the probe duration. (5) The timer and the work item added for this must be handed a count on net->servers_outstanding, which they hand on or release. This makes sure that network namespace cleanup waits for them. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Reported-by: Dave Botsch <botsch@cnf.cornell.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-24 22:10:00 +08:00
afs_server_trace_put_probe,
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
afs_server_trace_put_slist,
afs_server_trace_put_slist_isort,
afs_server_trace_put_uuid_rsq,
afs_server_trace_update,
};
afs: Add a tracepoint to track the lifetime of the afs_volume struct Add a tracepoint to track the lifetime of the afs_volume struct. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:02:04 +08:00
enum afs_volume_trace {
afs_volume_trace_alloc,
afs_volume_trace_free,
afs_volume_trace_get_alloc_sbi,
afs_volume_trace_get_cell_insert,
afs_volume_trace_get_new_op,
afs_volume_trace_get_query_alias,
afs_volume_trace_put_cell_dup,
afs_volume_trace_put_cell_root,
afs_volume_trace_put_destroy_sbi,
afs_volume_trace_put_free_fc,
afs_volume_trace_put_put_op,
afs_volume_trace_put_query_alias,
afs_volume_trace_put_validate_fc,
afs_volume_trace_remove,
};
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
enum afs_fs_operation {
afs_FS_FetchData = 130, /* AFS Fetch file data */
afs: Get an AFS3 ACL as an xattr Implement an xattr on AFS files called "afs.acl" that retrieves a file's ACL. It returns the raw AFS3 ACL from the result of calling FS.FetchACL, leaving any interpretation to userspace. Note that whilst YFS servers will respond to FS.FetchACL, this will render a more-advanced YFS ACL down. Use "afs.yfs.acl" instead for that. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:52 +08:00
afs_FS_FetchACL = 131, /* AFS Fetch file ACL */
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
afs_FS_FetchStatus = 132, /* AFS Fetch file status */
afs_FS_StoreData = 133, /* AFS Store file data */
afs: implement acl setting Implements the setting of ACLs in AFS by means of setting the afs.acl extended attribute on the file. Signed-off-by: Joe Gorse <jhgorse@gmail.com> Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:52 +08:00
afs_FS_StoreACL = 134, /* AFS Store file ACL */
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
afs_FS_StoreStatus = 135, /* AFS Store file status */
afs_FS_RemoveFile = 136, /* AFS Remove a file */
afs_FS_CreateFile = 137, /* AFS Create a file */
afs_FS_Rename = 138, /* AFS Rename or move a file or directory */
afs_FS_Symlink = 139, /* AFS Create a symbolic link */
afs_FS_Link = 140, /* AFS Create a hard link */
afs_FS_MakeDir = 141, /* AFS Create a directory */
afs_FS_RemoveDir = 142, /* AFS Remove a directory */
afs_FS_GetVolumeInfo = 148, /* AFS Get information about a volume */
afs_FS_GetVolumeStatus = 149, /* AFS Get volume status information */
afs_FS_GetRootVolume = 151, /* AFS Get root volume name */
afs_FS_SetLock = 156, /* AFS Request a file lock */
afs_FS_ExtendLock = 157, /* AFS Extend a file lock */
afs_FS_ReleaseLock = 158, /* AFS Release a file lock */
afs_FS_Lookup = 161, /* AFS lookup file in directory */
afs: Prospectively look up extra files when doing a single lookup When afs_lookup() is called, prospectively look up the next 50 uncached fids also from that same directory and cache the results, rather than just looking up the one file requested. This allows us to use the FS.InlineBulkStatus RPC op to increase efficiency by fetching up to 50 file statuses at a time. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-10 04:12:31 +08:00
afs_FS_InlineBulkStatus = 65536, /* AFS Fetch multiple file statuses with errors */
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
afs_FS_FetchData64 = 65537, /* AFS Fetch file data */
afs_FS_StoreData64 = 65538, /* AFS Store file data */
afs_FS_GiveUpAllCallBacks = 65539, /* AFS Give up all our callbacks on a server */
afs_FS_GetCapabilities = 65540, /* AFS Get FS server capabilities */
afs: Implement YFS support in the fs client Implement support for talking to YFS-variant fileservers in the cache manager and the filesystem client. These implement upgraded services on the same port as their AFS services. YFS fileservers provide expanded capabilities over AFS. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:58 +08:00
yfs_FS_FetchData = 130, /* YFS Fetch file data */
yfs_FS_FetchACL = 64131, /* YFS Fetch file ACL */
yfs_FS_FetchStatus = 64132, /* YFS Fetch file status */
yfs_FS_StoreACL = 64134, /* YFS Store file ACL */
yfs_FS_StoreStatus = 64135, /* YFS Store file status */
yfs_FS_RemoveFile = 64136, /* YFS Remove a file */
yfs_FS_CreateFile = 64137, /* YFS Create a file */
yfs_FS_Rename = 64138, /* YFS Rename or move a file or directory */
yfs_FS_Symlink = 64139, /* YFS Create a symbolic link */
yfs_FS_Link = 64140, /* YFS Create a hard link */
yfs_FS_MakeDir = 64141, /* YFS Create a directory */
yfs_FS_RemoveDir = 64142, /* YFS Remove a directory */
yfs_FS_GetVolumeStatus = 64149, /* YFS Get volume status information */
yfs_FS_SetVolumeStatus = 64150, /* YFS Set volume status information */
yfs_FS_SetLock = 64156, /* YFS Request a file lock */
yfs_FS_ExtendLock = 64157, /* YFS Extend a file lock */
yfs_FS_ReleaseLock = 64158, /* YFS Release a file lock */
yfs_FS_Lookup = 64161, /* YFS lookup file in directory */
yfs_FS_FlushCPS = 64165,
yfs_FS_FetchOpaqueACL = 64168,
yfs_FS_WhoAmI = 64170,
yfs_FS_RemoveACL = 64171,
yfs_FS_RemoveFile2 = 64173,
yfs_FS_StoreOpaqueACL2 = 64174,
yfs_FS_InlineBulkStatus = 64536, /* YFS Fetch multiple file statuses with errors */
yfs_FS_FetchData64 = 64537, /* YFS Fetch file data */
yfs_FS_StoreData64 = 64538, /* YFS Store file data */
yfs_FS_UpdateSymlink = 64540,
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
};
enum afs_vl_operation {
afs_VL_GetEntryByNameU = 527, /* AFS Get Vol Entry By Name operation ID */
afs_VL_GetAddrsU = 533, /* AFS Get FS server addresses */
afs_YFSVL_GetEndpoints = 64002, /* YFS Get FS & Vol server addresses */
afs: Implement client support for the YFSVL.GetCellName RPC op Implement client support for the YFSVL.GetCellName RPC operation by which YFS permits the canonical cell name to be queried from a VL server. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:26:41 +08:00
afs_YFSVL_GetCellName = 64014, /* YFS Get actual cell name */
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
afs_VL_GetCapabilities = 65537, /* AFS Get VL server capabilities */
};
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
enum afs_edit_dir_op {
afs_edit_dir_create,
afs_edit_dir_create_error,
afs_edit_dir_create_inval,
afs_edit_dir_create_nospc,
afs_edit_dir_delete,
afs_edit_dir_delete_error,
afs_edit_dir_delete_inval,
afs_edit_dir_delete_noent,
};
enum afs_edit_dir_reason {
afs_edit_dir_for_create,
afs_edit_dir_for_link,
afs_edit_dir_for_mkdir,
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
afs_edit_dir_for_rename_0,
afs_edit_dir_for_rename_1,
afs_edit_dir_for_rename_2,
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
afs_edit_dir_for_rmdir,
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
afs_edit_dir_for_silly_0,
afs_edit_dir_for_silly_1,
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
afs_edit_dir_for_symlink,
afs_edit_dir_for_unlink,
};
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
enum afs_eproto_cause {
afs_eproto_bad_status,
afs_eproto_cb_count,
afs_eproto_cb_fid_count,
afs: Implement client support for the YFSVL.GetCellName RPC op Implement client support for the YFSVL.GetCellName RPC operation by which YFS permits the canonical cell name to be queried from a VL server. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:26:41 +08:00
afs_eproto_cellname_len,
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
afs_eproto_file_type,
afs_eproto_ibulkst_cb_count,
afs_eproto_ibulkst_count,
afs_eproto_motd_len,
afs_eproto_offline_msg_len,
afs_eproto_volname_len,
afs_eproto_yvl_fsendpt4_len,
afs_eproto_yvl_fsendpt6_len,
afs_eproto_yvl_fsendpt_num,
afs_eproto_yvl_fsendpt_type,
afs_eproto_yvl_vlendpt4_len,
afs_eproto_yvl_vlendpt6_len,
afs_eproto_yvl_vlendpt_type,
};
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
enum afs_io_error {
afs_io_error_cm_reply,
afs_io_error_extract,
afs_io_error_fs_probe_fail,
afs_io_error_vl_lookup_fail,
afs: Probe multiple fileservers simultaneously Send probes to all the unprobed fileservers in a fileserver list on all addresses simultaneously in an attempt to find out the fastest route whilst not getting stuck for 20s on any server or address that we don't get a reply from. This alleviates the problem whereby attempting to access a new server can take a long time because the rotation algorithm ends up rotating through all servers and addresses until it finds one that responds. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:59 +08:00
afs_io_error_vl_probe_fail,
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
};
enum afs_file_error {
afs_file_error_dir_bad_magic,
afs_file_error_dir_big,
afs_file_error_dir_missing_page,
afs_file_error_dir_over_end,
afs_file_error_dir_small,
afs_file_error_dir_unmarked_ext,
afs_file_error_mntpt,
afs_file_error_writeback_fail,
};
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
enum afs_flock_event {
afs_flock_acquired,
afs_flock_callback_break,
afs_flock_defer_unlock,
afs: Handle lock rpc ops failing on a file that got deleted Holding a file lock on an AFS file does not prevent it from being deleted on the server, so we need to handle an error resulting from that when we try setting, extending or releasing a lock. Fix this by adding a "deleted" lock state and cancelling the lock extension process for that file and aborting all waiters for the lock. Fixes: 0fafdc9f888b ("afs: Fix file locking") Reported-by: Jonathan Billings <jsbillin@umich.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
afs_flock_extend_fail,
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
afs_flock_fail_other,
afs_flock_fail_perm,
afs_flock_no_lockers,
afs: Handle lock rpc ops failing on a file that got deleted Holding a file lock on an AFS file does not prevent it from being deleted on the server, so we need to handle an error resulting from that when we try setting, extending or releasing a lock. Fix this by adding a "deleted" lock state and cancelling the lock extension process for that file and aborting all waiters for the lock. Fixes: 0fafdc9f888b ("afs: Fix file locking") Reported-by: Jonathan Billings <jsbillin@umich.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
afs_flock_release_fail,
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
afs_flock_silly_delete,
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
afs_flock_timestamp,
afs_flock_try_to_lock,
afs_flock_vfs_lock,
afs_flock_vfs_locking,
afs_flock_waited,
afs_flock_waiting,
afs_flock_work_extending,
afs_flock_work_retry,
afs_flock_work_unlocking,
afs_flock_would_block,
};
enum afs_flock_operation {
afs_flock_op_copy_lock,
afs_flock_op_flock,
afs_flock_op_grant,
afs_flock_op_lock,
afs_flock_op_release_lock,
afs_flock_op_return_ok,
afs_flock_op_return_eagain,
afs_flock_op_return_edeadlk,
afs_flock_op_return_error,
afs_flock_op_set_lock,
afs_flock_op_unlock,
afs_flock_op_wake,
};
afs: Add some callback management tracepoints Add a couple of tracepoints to track callback management: (1) afs_cb_miss - Logs when we were unable to apply a callback, either due to the inode being discarded or due to a competing thread applying a callback first. (2) afs_cb_break - Logs when we attempted to clear the noted callback promise, either due to the server explicitly breaking the callback, the callback promise lapsing or a local event obsoleting it. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:16 +08:00
enum afs_cb_break_reason {
afs_cb_break_no_break,
afs_cb_break_for_callback,
afs_cb_break_for_deleted,
afs_cb_break_for_lapsed,
afs_cb_break_for_unlink,
afs_cb_break_for_vsbreak,
afs_cb_break_for_volume_callback,
afs_cb_break_for_zap,
};
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
#endif /* end __AFS_DECLARE_TRACE_ENUMS_ONCE_ONLY */
/*
* Declare tracing information enums and their string mappings for display.
*/
#define afs_call_traces \
EM(afs_call_trace_alloc, "ALLOC") \
EM(afs_call_trace_free, "FREE ") \
afs: Fix race in async call refcounting There's a race between afs_make_call() and afs_wake_up_async_call() in the case that an error is returned from rxrpc_kernel_send_data() after it has queued the final packet. afs_make_call() will try and clean up the mess, but the call state may have been moved on thereby causing afs_process_async_call() to also try and to delete the call. Fix this by: (1) Getting an extra ref for an asynchronous call for the call itself to hold. This makes sure the call doesn't evaporate on us accidentally and will allow the call to be retained by the caller in a future patch. The ref is released on leaving afs_make_call() or afs_wait_for_call_to_complete(). (2) In the event of an error from rxrpc_kernel_send_data(): (a) Don't set the call state to AFS_CALL_COMPLETE until *after* the call has been aborted and ended. This prevents afs_deliver_to_call() from doing anything with any notifications it gets. (b) Explicitly end the call immediately to prevent further callbacks. (c) Cancel any queued async_work and wait for the work if it's executing. This allows us to be sure the race won't recur when we change the state. We put the work queue's ref on the call if we managed to cancel it. (d) Put the call's ref that we got in (1). This belongs to us as long as the call is in state AFS_CALL_CL_REQUESTING. Fixes: 341f741f04be ("afs: Refcount the afs_call struct") Signed-off-by: David Howells <dhowells@redhat.com>
2019-01-10 23:40:50 +08:00
EM(afs_call_trace_get, "GET ") \
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
EM(afs_call_trace_put, "PUT ") \
EM(afs_call_trace_wake, "WAKE ") \
afs: Fix some tracing details Fix a couple of tracelines to indicate the usage count after the atomic op, not the usage count before it to be consistent with other afs and rxrpc trace lines. Change the wording of the afs_call_trace_work trace ID label from "WORK" to "QUEUE" to reflect the fact that it's queueing work, not doing work. Fixes: 341f741f04be ("afs: Refcount the afs_call struct") Signed-off-by: David Howells <dhowells@redhat.com>
2020-03-13 21:36:01 +08:00
E_(afs_call_trace_work, "QUEUE")
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
#define afs_server_traces \
EM(afs_server_trace_alloc, "ALLOC ") \
EM(afs_server_trace_callback, "CALLBACK ") \
EM(afs_server_trace_destroy, "DESTROY ") \
EM(afs_server_trace_free, "FREE ") \
EM(afs_server_trace_gc, "GC ") \
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
EM(afs_server_trace_get_by_addr, "GET addr ") \
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
EM(afs_server_trace_get_by_uuid, "GET uuid ") \
EM(afs_server_trace_get_caps, "GET caps ") \
EM(afs_server_trace_get_install, "GET inst ") \
EM(afs_server_trace_get_new_cbi, "GET cbi ") \
afs: Actively poll fileservers to maintain NAT or firewall openings When an AFS client accesses a file, it receives a limited-duration callback promise that the server will notify it if another client changes a file. This callback duration can be a few hours in length. If a client mounts a volume and then an application prevents it from being unmounted, say by chdir'ing into it, but then does nothing for some time, the rxrpc_peer record will expire and rxrpc-level keepalive will cease. If there is NAT or a firewall between the client and the server, the route back for the server may close after a comparatively short duration, meaning that attempts by the server to notify the client may then bounce. The client, however, may (so far as it knows) still have a valid unexpired promise and will then rely on its cached data and will not see changes made on the server by a third party until it incidentally rechecks the status or the promise needs renewal. To deal with this, the client needs to regularly probe the server. This has two effects: firstly, it keeps a route open back for the server, and secondly, it causes the server to disgorge any notifications that got queued up because they couldn't be sent. Fix this by adding a mechanism to emit regular probes. Two levels of probing are made available: Under normal circumstances the 'slow' queue will be used for a fileserver - this just probes the preferred address once every 5 mins or so; however, if server fails to respond to any probes, the server will shift to the 'fast' queue from which all its interfaces will be probed every 30s. When it finally responds, the record will switch back to the slow queue. Further notes: (1) Probing is now no longer driven from the fileserver rotation algorithm. (2) Probes are dispatched to all interfaces on a fileserver when that an afs_server object is set up to record it. (3) The afs_server object is removed from the probe queues when we start to probe it. afs_is_probing_server() returns true if it's not listed - ie. it's undergoing probing. (4) The afs_server object is added back on to the probe queue when the final outstanding probe completes, but the probed_at time is set when we're about to launch a probe so that it's not dependent on the probe duration. (5) The timer and the work item added for this must be handed a count on net->servers_outstanding, which they hand on or release. This makes sure that network namespace cleanup waits for them. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Reported-by: Dave Botsch <botsch@cnf.cornell.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-24 22:10:00 +08:00
EM(afs_server_trace_get_probe, "GET probe") \
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
EM(afs_server_trace_give_up_cb, "giveup-cb") \
EM(afs_server_trace_put_call, "PUT call ") \
EM(afs_server_trace_put_cbi, "PUT cbi ") \
EM(afs_server_trace_put_find_rsq, "PUT f-rsq") \
afs: Actively poll fileservers to maintain NAT or firewall openings When an AFS client accesses a file, it receives a limited-duration callback promise that the server will notify it if another client changes a file. This callback duration can be a few hours in length. If a client mounts a volume and then an application prevents it from being unmounted, say by chdir'ing into it, but then does nothing for some time, the rxrpc_peer record will expire and rxrpc-level keepalive will cease. If there is NAT or a firewall between the client and the server, the route back for the server may close after a comparatively short duration, meaning that attempts by the server to notify the client may then bounce. The client, however, may (so far as it knows) still have a valid unexpired promise and will then rely on its cached data and will not see changes made on the server by a third party until it incidentally rechecks the status or the promise needs renewal. To deal with this, the client needs to regularly probe the server. This has two effects: firstly, it keeps a route open back for the server, and secondly, it causes the server to disgorge any notifications that got queued up because they couldn't be sent. Fix this by adding a mechanism to emit regular probes. Two levels of probing are made available: Under normal circumstances the 'slow' queue will be used for a fileserver - this just probes the preferred address once every 5 mins or so; however, if server fails to respond to any probes, the server will shift to the 'fast' queue from which all its interfaces will be probed every 30s. When it finally responds, the record will switch back to the slow queue. Further notes: (1) Probing is now no longer driven from the fileserver rotation algorithm. (2) Probes are dispatched to all interfaces on a fileserver when that an afs_server object is set up to record it. (3) The afs_server object is removed from the probe queues when we start to probe it. afs_is_probing_server() returns true if it's not listed - ie. it's undergoing probing. (4) The afs_server object is added back on to the probe queue when the final outstanding probe completes, but the probed_at time is set when we're about to launch a probe so that it's not dependent on the probe duration. (5) The timer and the work item added for this must be handed a count on net->servers_outstanding, which they hand on or release. This makes sure that network namespace cleanup waits for them. Fixes: d2ddc776a458 ("afs: Overhaul volume and server record caching and fileserver rotation") Reported-by: Dave Botsch <botsch@cnf.cornell.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-24 22:10:00 +08:00
EM(afs_server_trace_put_probe, "PUT probe") \
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
EM(afs_server_trace_put_slist, "PUT slist") \
EM(afs_server_trace_put_slist_isort, "PUT isort") \
EM(afs_server_trace_put_uuid_rsq, "PUT u-req") \
E_(afs_server_trace_update, "UPDATE")
afs: Add a tracepoint to track the lifetime of the afs_volume struct Add a tracepoint to track the lifetime of the afs_volume struct. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:02:04 +08:00
#define afs_volume_traces \
EM(afs_volume_trace_alloc, "ALLOC ") \
EM(afs_volume_trace_free, "FREE ") \
EM(afs_volume_trace_get_alloc_sbi, "GET sbi-alloc ") \
EM(afs_volume_trace_get_cell_insert, "GET cell-insrt") \
EM(afs_volume_trace_get_new_op, "GET op-new ") \
EM(afs_volume_trace_get_query_alias, "GET cell-alias") \
EM(afs_volume_trace_put_cell_dup, "PUT cell-dup ") \
EM(afs_volume_trace_put_cell_root, "PUT cell-root ") \
EM(afs_volume_trace_put_destroy_sbi, "PUT sbi-destry") \
EM(afs_volume_trace_put_free_fc, "PUT fc-free ") \
EM(afs_volume_trace_put_put_op, "PUT op-put ") \
EM(afs_volume_trace_put_query_alias, "PUT cell-alias") \
EM(afs_volume_trace_put_validate_fc, "PUT fc-validat") \
E_(afs_volume_trace_remove, "REMOVE ")
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
#define afs_fs_operations \
EM(afs_FS_FetchData, "FS.FetchData") \
EM(afs_FS_FetchStatus, "FS.FetchStatus") \
EM(afs_FS_StoreData, "FS.StoreData") \
EM(afs_FS_StoreStatus, "FS.StoreStatus") \
EM(afs_FS_RemoveFile, "FS.RemoveFile") \
EM(afs_FS_CreateFile, "FS.CreateFile") \
EM(afs_FS_Rename, "FS.Rename") \
EM(afs_FS_Symlink, "FS.Symlink") \
EM(afs_FS_Link, "FS.Link") \
EM(afs_FS_MakeDir, "FS.MakeDir") \
EM(afs_FS_RemoveDir, "FS.RemoveDir") \
EM(afs_FS_GetVolumeInfo, "FS.GetVolumeInfo") \
EM(afs_FS_GetVolumeStatus, "FS.GetVolumeStatus") \
EM(afs_FS_GetRootVolume, "FS.GetRootVolume") \
EM(afs_FS_SetLock, "FS.SetLock") \
EM(afs_FS_ExtendLock, "FS.ExtendLock") \
EM(afs_FS_ReleaseLock, "FS.ReleaseLock") \
EM(afs_FS_Lookup, "FS.Lookup") \
afs: Prospectively look up extra files when doing a single lookup When afs_lookup() is called, prospectively look up the next 50 uncached fids also from that same directory and cache the results, rather than just looking up the one file requested. This allows us to use the FS.InlineBulkStatus RPC op to increase efficiency by fetching up to 50 file statuses at a time. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-10 04:12:31 +08:00
EM(afs_FS_InlineBulkStatus, "FS.InlineBulkStatus") \
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
EM(afs_FS_FetchData64, "FS.FetchData64") \
EM(afs_FS_StoreData64, "FS.StoreData64") \
EM(afs_FS_GiveUpAllCallBacks, "FS.GiveUpAllCallBacks") \
afs: Implement YFS support in the fs client Implement support for talking to YFS-variant fileservers in the cache manager and the filesystem client. These implement upgraded services on the same port as their AFS services. YFS fileservers provide expanded capabilities over AFS. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:58 +08:00
EM(afs_FS_GetCapabilities, "FS.GetCapabilities") \
EM(yfs_FS_FetchACL, "YFS.FetchACL") \
EM(yfs_FS_FetchStatus, "YFS.FetchStatus") \
EM(yfs_FS_StoreACL, "YFS.StoreACL") \
EM(yfs_FS_StoreStatus, "YFS.StoreStatus") \
EM(yfs_FS_RemoveFile, "YFS.RemoveFile") \
EM(yfs_FS_CreateFile, "YFS.CreateFile") \
EM(yfs_FS_Rename, "YFS.Rename") \
EM(yfs_FS_Symlink, "YFS.Symlink") \
EM(yfs_FS_Link, "YFS.Link") \
EM(yfs_FS_MakeDir, "YFS.MakeDir") \
EM(yfs_FS_RemoveDir, "YFS.RemoveDir") \
EM(yfs_FS_GetVolumeStatus, "YFS.GetVolumeStatus") \
EM(yfs_FS_SetVolumeStatus, "YFS.SetVolumeStatus") \
EM(yfs_FS_SetLock, "YFS.SetLock") \
EM(yfs_FS_ExtendLock, "YFS.ExtendLock") \
EM(yfs_FS_ReleaseLock, "YFS.ReleaseLock") \
EM(yfs_FS_Lookup, "YFS.Lookup") \
EM(yfs_FS_FlushCPS, "YFS.FlushCPS") \
EM(yfs_FS_FetchOpaqueACL, "YFS.FetchOpaqueACL") \
EM(yfs_FS_WhoAmI, "YFS.WhoAmI") \
EM(yfs_FS_RemoveACL, "YFS.RemoveACL") \
EM(yfs_FS_RemoveFile2, "YFS.RemoveFile2") \
EM(yfs_FS_StoreOpaqueACL2, "YFS.StoreOpaqueACL2") \
EM(yfs_FS_InlineBulkStatus, "YFS.InlineBulkStatus") \
EM(yfs_FS_FetchData64, "YFS.FetchData64") \
EM(yfs_FS_StoreData64, "YFS.StoreData64") \
E_(yfs_FS_UpdateSymlink, "YFS.UpdateSymlink")
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
#define afs_vl_operations \
EM(afs_VL_GetEntryByNameU, "VL.GetEntryByNameU") \
EM(afs_VL_GetAddrsU, "VL.GetAddrsU") \
EM(afs_YFSVL_GetEndpoints, "YFSVL.GetEndpoints") \
afs: Implement client support for the YFSVL.GetCellName RPC op Implement client support for the YFSVL.GetCellName RPC operation by which YFS permits the canonical cell name to be queried from a VL server. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:26:41 +08:00
EM(afs_YFSVL_GetCellName, "YFSVL.GetCellName") \
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
E_(afs_VL_GetCapabilities, "VL.GetCapabilities")
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
#define afs_edit_dir_ops \
EM(afs_edit_dir_create, "create") \
EM(afs_edit_dir_create_error, "c_fail") \
EM(afs_edit_dir_create_inval, "c_invl") \
EM(afs_edit_dir_create_nospc, "c_nspc") \
EM(afs_edit_dir_delete, "delete") \
EM(afs_edit_dir_delete_error, "d_err ") \
EM(afs_edit_dir_delete_inval, "d_invl") \
E_(afs_edit_dir_delete_noent, "d_nent")
#define afs_edit_dir_reasons \
EM(afs_edit_dir_for_create, "Create") \
EM(afs_edit_dir_for_link, "Link ") \
EM(afs_edit_dir_for_mkdir, "MkDir ") \
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
EM(afs_edit_dir_for_rename_0, "Renam0") \
EM(afs_edit_dir_for_rename_1, "Renam1") \
EM(afs_edit_dir_for_rename_2, "Renam2") \
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
EM(afs_edit_dir_for_rmdir, "RmDir ") \
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
EM(afs_edit_dir_for_silly_0, "S_Ren0") \
EM(afs_edit_dir_for_silly_1, "S_Ren1") \
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
EM(afs_edit_dir_for_symlink, "Symlnk") \
E_(afs_edit_dir_for_unlink, "Unlink")
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
#define afs_eproto_causes \
EM(afs_eproto_bad_status, "BadStatus") \
EM(afs_eproto_cb_count, "CbCount") \
EM(afs_eproto_cb_fid_count, "CbFidCount") \
afs: Implement client support for the YFSVL.GetCellName RPC op Implement client support for the YFSVL.GetCellName RPC operation by which YFS permits the canonical cell name to be queried from a VL server. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:26:41 +08:00
EM(afs_eproto_cellname_len, "CellNameLen") \
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
EM(afs_eproto_file_type, "FileTYpe") \
EM(afs_eproto_ibulkst_cb_count, "IBS.CbCount") \
EM(afs_eproto_ibulkst_count, "IBS.FidCount") \
EM(afs_eproto_motd_len, "MotdLen") \
EM(afs_eproto_offline_msg_len, "OfflineMsgLen") \
EM(afs_eproto_volname_len, "VolNameLen") \
EM(afs_eproto_yvl_fsendpt4_len, "YVL.FsEnd4Len") \
EM(afs_eproto_yvl_fsendpt6_len, "YVL.FsEnd6Len") \
EM(afs_eproto_yvl_fsendpt_num, "YVL.FsEndCount") \
EM(afs_eproto_yvl_fsendpt_type, "YVL.FsEndType") \
EM(afs_eproto_yvl_vlendpt4_len, "YVL.VlEnd4Len") \
EM(afs_eproto_yvl_vlendpt6_len, "YVL.VlEnd6Len") \
E_(afs_eproto_yvl_vlendpt_type, "YVL.VlEndType")
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
#define afs_io_errors \
EM(afs_io_error_cm_reply, "CM_REPLY") \
EM(afs_io_error_extract, "EXTRACT") \
EM(afs_io_error_fs_probe_fail, "FS_PROBE_FAIL") \
afs: Probe multiple fileservers simultaneously Send probes to all the unprobed fileservers in a fileserver list on all addresses simultaneously in an attempt to find out the fastest route whilst not getting stuck for 20s on any server or address that we don't get a reply from. This alleviates the problem whereby attempting to access a new server can take a long time because the rotation algorithm ends up rotating through all servers and addresses until it finds one that responds. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:59 +08:00
EM(afs_io_error_vl_lookup_fail, "VL_LOOKUP_FAIL") \
E_(afs_io_error_vl_probe_fail, "VL_PROBE_FAIL")
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
#define afs_file_errors \
EM(afs_file_error_dir_bad_magic, "DIR_BAD_MAGIC") \
EM(afs_file_error_dir_big, "DIR_BIG") \
EM(afs_file_error_dir_missing_page, "DIR_MISSING_PAGE") \
EM(afs_file_error_dir_over_end, "DIR_ENT_OVER_END") \
EM(afs_file_error_dir_small, "DIR_SMALL") \
EM(afs_file_error_dir_unmarked_ext, "DIR_UNMARKED_EXT") \
EM(afs_file_error_mntpt, "MNTPT_READ_FAILED") \
E_(afs_file_error_writeback_fail, "WRITEBACK_FAILED")
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
#define afs_flock_types \
EM(F_RDLCK, "RDLCK") \
EM(F_WRLCK, "WRLCK") \
E_(F_UNLCK, "UNLCK")
#define afs_flock_states \
EM(AFS_VNODE_LOCK_NONE, "NONE") \
EM(AFS_VNODE_LOCK_WAITING_FOR_CB, "WAIT_FOR_CB") \
EM(AFS_VNODE_LOCK_SETTING, "SETTING") \
EM(AFS_VNODE_LOCK_GRANTED, "GRANTED") \
EM(AFS_VNODE_LOCK_EXTENDING, "EXTENDING") \
EM(AFS_VNODE_LOCK_NEED_UNLOCK, "NEED_UNLOCK") \
afs: Handle lock rpc ops failing on a file that got deleted Holding a file lock on an AFS file does not prevent it from being deleted on the server, so we need to handle an error resulting from that when we try setting, extending or releasing a lock. Fix this by adding a "deleted" lock state and cancelling the lock extension process for that file and aborting all waiters for the lock. Fixes: 0fafdc9f888b ("afs: Fix file locking") Reported-by: Jonathan Billings <jsbillin@umich.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
EM(AFS_VNODE_LOCK_UNLOCKING, "UNLOCKING") \
E_(AFS_VNODE_LOCK_DELETED, "DELETED")
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
#define afs_flock_events \
EM(afs_flock_acquired, "Acquired") \
EM(afs_flock_callback_break, "Callback") \
EM(afs_flock_defer_unlock, "D-Unlock") \
afs: Handle lock rpc ops failing on a file that got deleted Holding a file lock on an AFS file does not prevent it from being deleted on the server, so we need to handle an error resulting from that when we try setting, extending or releasing a lock. Fix this by adding a "deleted" lock state and cancelling the lock extension process for that file and aborting all waiters for the lock. Fixes: 0fafdc9f888b ("afs: Fix file locking") Reported-by: Jonathan Billings <jsbillin@umich.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
EM(afs_flock_extend_fail, "Ext_Fail") \
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
EM(afs_flock_fail_other, "ErrOther") \
EM(afs_flock_fail_perm, "ErrPerm ") \
EM(afs_flock_no_lockers, "NoLocker") \
afs: Handle lock rpc ops failing on a file that got deleted Holding a file lock on an AFS file does not prevent it from being deleted on the server, so we need to handle an error resulting from that when we try setting, extending or releasing a lock. Fix this by adding a "deleted" lock state and cancelling the lock extension process for that file and aborting all waiters for the lock. Fixes: 0fafdc9f888b ("afs: Fix file locking") Reported-by: Jonathan Billings <jsbillin@umich.edu> Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
EM(afs_flock_release_fail, "Rel_Fail") \
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
EM(afs_flock_silly_delete, "SillyDel") \
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
EM(afs_flock_timestamp, "Timestmp") \
EM(afs_flock_try_to_lock, "TryToLck") \
EM(afs_flock_vfs_lock, "VFSLock ") \
EM(afs_flock_vfs_locking, "VFSLking") \
EM(afs_flock_waited, "Waited ") \
EM(afs_flock_waiting, "Waiting ") \
EM(afs_flock_work_extending, "Extendng") \
EM(afs_flock_work_retry, "Retry ") \
EM(afs_flock_work_unlocking, "Unlcking") \
E_(afs_flock_would_block, "EWOULDBL")
#define afs_flock_operations \
EM(afs_flock_op_copy_lock, "COPY ") \
EM(afs_flock_op_flock, "->flock ") \
EM(afs_flock_op_grant, "GRANT ") \
EM(afs_flock_op_lock, "->lock ") \
EM(afs_flock_op_release_lock, "RELEASE ") \
EM(afs_flock_op_return_ok, "<-OK ") \
EM(afs_flock_op_return_edeadlk, "<-EDEADL") \
EM(afs_flock_op_return_eagain, "<-EAGAIN") \
EM(afs_flock_op_return_error, "<-ERROR ") \
EM(afs_flock_op_set_lock, "SET ") \
EM(afs_flock_op_unlock, "UNLOCK ") \
E_(afs_flock_op_wake, "WAKE ")
afs: Add some callback management tracepoints Add a couple of tracepoints to track callback management: (1) afs_cb_miss - Logs when we were unable to apply a callback, either due to the inode being discarded or due to a competing thread applying a callback first. (2) afs_cb_break - Logs when we attempted to clear the noted callback promise, either due to the server explicitly breaking the callback, the callback promise lapsing or a local event obsoleting it. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:16 +08:00
#define afs_cb_break_reasons \
EM(afs_cb_break_no_break, "no-break") \
EM(afs_cb_break_for_callback, "break-cb") \
EM(afs_cb_break_for_deleted, "break-del") \
EM(afs_cb_break_for_lapsed, "break-lapsed") \
EM(afs_cb_break_for_unlink, "break-unlink") \
EM(afs_cb_break_for_vsbreak, "break-vs") \
EM(afs_cb_break_for_volume_callback, "break-v-cb") \
E_(afs_cb_break_for_zap, "break-zap")
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
/*
* Export enum symbols via userspace.
*/
#undef EM
#undef E_
#define EM(a, b) TRACE_DEFINE_ENUM(a);
#define E_(a, b) TRACE_DEFINE_ENUM(a);
afs_call_traces;
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
afs_server_traces;
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
afs_fs_operations;
afs_vl_operations;
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
afs_edit_dir_ops;
afs_edit_dir_reasons;
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
afs_eproto_causes;
afs_io_errors;
afs_file_errors;
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
afs_flock_types;
afs_flock_operations;
afs: Add some callback management tracepoints Add a couple of tracepoints to track callback management: (1) afs_cb_miss - Logs when we were unable to apply a callback, either due to the inode being discarded or due to a competing thread applying a callback first. (2) afs_cb_break - Logs when we attempted to clear the noted callback promise, either due to the server explicitly breaking the callback, the callback promise lapsing or a local event obsoleting it. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:16 +08:00
afs_cb_break_reasons;
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
/*
* Now redefine the EM() and E_() macros to map the enums to the strings that
* will be printed in the output.
*/
#undef EM
#undef E_
#define EM(a, b) { a, b },
#define E_(a, b) { a, b }
afs: Set up the iov_iter before calling afs_extract_data() afs_extract_data sets up a temporary iov_iter and passes it to AF_RXRPC each time it is called to describe the remaining buffer to be filled. Instead: (1) Put an iterator in the afs_call struct. (2) Set the iterator for each marshalling stage to load data into the appropriate places. A number of convenience functions are provided to this end (eg. afs_extract_to_buf()). This iterator is then passed to afs_extract_data(). (3) Use the new ITER_DISCARD iterator to discard any excess data provided by FetchData. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
TRACE_EVENT(afs_receive_data,
TP_PROTO(struct afs_call *call, struct iov_iter *iter,
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
bool want_more, int ret),
afs: Set up the iov_iter before calling afs_extract_data() afs_extract_data sets up a temporary iov_iter and passes it to AF_RXRPC each time it is called to describe the remaining buffer to be filled. Instead: (1) Put an iterator in the afs_call struct. (2) Set the iterator for each marshalling stage to load data into the appropriate places. A number of convenience functions are provided to this end (eg. afs_extract_to_buf()). This iterator is then passed to afs_extract_data(). (3) Use the new ITER_DISCARD iterator to discard any excess data provided by FetchData. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
TP_ARGS(call, iter, want_more, ret),
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
TP_STRUCT__entry(
afs: Set up the iov_iter before calling afs_extract_data() afs_extract_data sets up a temporary iov_iter and passes it to AF_RXRPC each time it is called to describe the remaining buffer to be filled. Instead: (1) Put an iterator in the afs_call struct. (2) Set the iterator for each marshalling stage to load data into the appropriate places. A number of convenience functions are provided to this end (eg. afs_extract_to_buf()). This iterator is then passed to afs_extract_data(). (3) Use the new ITER_DISCARD iterator to discard any excess data provided by FetchData. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
__field(loff_t, remain )
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__field(enum afs_call_state, state )
__field(unsigned short, unmarshall )
__field(bool, want_more )
__field(int, ret )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->state = call->state;
__entry->unmarshall = call->unmarshall;
afs: Set up the iov_iter before calling afs_extract_data() afs_extract_data sets up a temporary iov_iter and passes it to AF_RXRPC each time it is called to describe the remaining buffer to be filled. Instead: (1) Put an iterator in the afs_call struct. (2) Set the iterator for each marshalling stage to load data into the appropriate places. A number of convenience functions are provided to this end (eg. afs_extract_to_buf()). This iterator is then passed to afs_extract_data(). (3) Use the new ITER_DISCARD iterator to discard any excess data provided by FetchData. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
__entry->remain = iov_iter_count(iter);
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->want_more = want_more;
__entry->ret = ret;
),
afs: Set up the iov_iter before calling afs_extract_data() afs_extract_data sets up a temporary iov_iter and passes it to AF_RXRPC each time it is called to describe the remaining buffer to be filled. Instead: (1) Put an iterator in the afs_call struct. (2) Set the iterator for each marshalling stage to load data into the appropriate places. A number of convenience functions are provided to this end (eg. afs_extract_to_buf()). This iterator is then passed to afs_extract_data(). (3) Use the new ITER_DISCARD iterator to discard any excess data provided by FetchData. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
TP_printk("c=%08x r=%llu u=%u w=%u s=%u ret=%d",
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->call,
afs: Set up the iov_iter before calling afs_extract_data() afs_extract_data sets up a temporary iov_iter and passes it to AF_RXRPC each time it is called to describe the remaining buffer to be filled. Instead: (1) Put an iterator in the afs_call struct. (2) Set the iterator for each marshalling stage to load data into the appropriate places. A number of convenience functions are provided to this end (eg. afs_extract_to_buf()). This iterator is then passed to afs_extract_data(). (3) Use the new ITER_DISCARD iterator to discard any excess data provided by FetchData. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
__entry->remain,
__entry->unmarshall,
__entry->want_more,
__entry->state,
__entry->ret)
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
);
TRACE_EVENT(afs_notify_call,
TP_PROTO(struct rxrpc_call *rxcall, struct afs_call *call),
TP_ARGS(rxcall, call),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__field(enum afs_call_state, state )
__field(unsigned short, unmarshall )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->state = call->state;
__entry->unmarshall = call->unmarshall;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk("c=%08x s=%u u=%u",
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->call,
__entry->state, __entry->unmarshall)
);
TRACE_EVENT(afs_cb_call,
TP_PROTO(struct afs_call *call),
TP_ARGS(call),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__field(const char *, name )
__field(u32, op )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->name = call->type->name;
__entry->op = call->operation_ID;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk("c=%08x %s o=%u",
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
__entry->call,
__entry->name,
__entry->op)
);
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
TRACE_EVENT(afs_call,
TP_PROTO(struct afs_call *call, enum afs_call_trace op,
int usage, int outstanding, const void *where),
TP_ARGS(call, op, usage, outstanding, where),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
__field(int, op )
__field(int, usage )
__field(int, outstanding )
__field(const void *, where )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
__entry->op = op;
__entry->usage = usage;
__entry->outstanding = outstanding;
__entry->where = where;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk("c=%08x %s u=%d o=%d sp=%pSR",
afs: Refcount the afs_call struct A static checker warning occurs in the AFS filesystem: fs/afs/cmservice.c:155 SRXAFSCB_CallBack() error: dereferencing freed memory 'call' due to the reply being sent before we access the server it points to. The act of sending the reply causes the call to be freed if an error occurs (but not if it doesn't). On top of this, the lifetime handling of afs_call structs is fragile because they get passed around through workqueues without any sort of refcounting. Deal with the issues by: (1) Fix the maybe/maybe not nature of the reply sending functions with regards to whether they release the call struct. (2) Refcount the afs_call struct and sort out places that need to get/put references. (3) Pass a ref through the work queue and release (or pass on) that ref in the work function. Care has to be taken because a work queue may already own a ref to the call. (4) Do the cleaning up in the put function only. (5) Simplify module cleanup by always incrementing afs_outstanding_calls whenever a call is allocated. (6) Set the backlog to 0 with kernel_listen() at the beginning of the process of closing the socket to prevent new incoming calls from occurring and to remove the contribution of preallocated calls from afs_outstanding_calls before we wait on it. A tracepoint is also added to monitor the afs_call refcount and lifetime. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Fixes: 08e0e7c82eea: "[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC."
2017-01-05 18:38:36 +08:00
__entry->call,
__print_symbolic(__entry->op, afs_call_traces),
__entry->usage,
__entry->outstanding,
__entry->where)
);
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
TRACE_EVENT(afs_make_fs_call,
TP_PROTO(struct afs_call *call, const struct afs_fid *fid),
TP_ARGS(call, fid),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__field(enum afs_fs_operation, op )
__field_struct(struct afs_fid, fid )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->op = call->operation_ID;
if (fid) {
__entry->fid = *fid;
} else {
__entry->fid.vid = 0;
__entry->fid.vnode = 0;
__entry->fid.unique = 0;
}
),
afs: Increase to 64-bit volume ID and 96-bit vnode ID for YFS Increase the sizes of the volume ID to 64 bits and the vnode ID (inode number equivalent) to 96 bits to allow the support of YFS. This requires the iget comparator to check the vnode->fid rather than i_ino and i_generation as i_ino is not sufficiently capacious. It also requires this data to be placed into the vnode cache key for fscache. For the moment, just discard the top 32 bits of the vnode ID when returning it though stat. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
TP_printk("c=%08x %06llx:%06llx:%06x %s",
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->call,
__entry->fid.vid,
__entry->fid.vnode,
__entry->fid.unique,
__print_symbolic(__entry->op, afs_fs_operations))
);
afs: Provide mount-time configurable byte-range file locking emulation Provide byte-range file locking emulation that can be configured at mount time to one of four modes: (1) flock=local. Locking is done locally only and no reference is made to the server. (2) flock=openafs. Byte-range locking is done locally only; whole-file locking is done with reference to the server. Whole-file locks cannot be upgraded unless the client holds an exclusive lock. (3) flock=strict. Byte-range and whole-file locking both require a sufficient whole-file lock on the server. (4) flock=write. As strict, but the client always gets an exclusive whole-file lock on the server. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:52 +08:00
TRACE_EVENT(afs_make_fs_calli,
TP_PROTO(struct afs_call *call, const struct afs_fid *fid,
unsigned int i),
TP_ARGS(call, fid, i),
TP_STRUCT__entry(
__field(unsigned int, call )
__field(unsigned int, i )
__field(enum afs_fs_operation, op )
__field_struct(struct afs_fid, fid )
),
TP_fast_assign(
__entry->call = call->debug_id;
__entry->i = i;
__entry->op = call->operation_ID;
if (fid) {
__entry->fid = *fid;
} else {
__entry->fid.vid = 0;
__entry->fid.vnode = 0;
__entry->fid.unique = 0;
}
),
TP_printk("c=%08x %06llx:%06llx:%06x %s i=%u",
__entry->call,
__entry->fid.vid,
__entry->fid.vnode,
__entry->fid.unique,
__print_symbolic(__entry->op, afs_fs_operations),
__entry->i)
);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TRACE_EVENT(afs_make_fs_call1,
TP_PROTO(struct afs_call *call, const struct afs_fid *fid,
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
const struct qstr *name),
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TP_ARGS(call, fid, name),
TP_STRUCT__entry(
__field(unsigned int, call )
__field(enum afs_fs_operation, op )
__field_struct(struct afs_fid, fid )
__array(char, name, 24 )
),
TP_fast_assign(
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
unsigned int __len = min_t(unsigned int, name->len, 23);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__entry->call = call->debug_id;
__entry->op = call->operation_ID;
if (fid) {
__entry->fid = *fid;
} else {
__entry->fid.vid = 0;
__entry->fid.vnode = 0;
__entry->fid.unique = 0;
}
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
memcpy(__entry->name, name->name, __len);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__entry->name[__len] = 0;
),
TP_printk("c=%08x %06llx:%06llx:%06x %s \"%s\"",
__entry->call,
__entry->fid.vid,
__entry->fid.vnode,
__entry->fid.unique,
__print_symbolic(__entry->op, afs_fs_operations),
__entry->name)
);
TRACE_EVENT(afs_make_fs_call2,
TP_PROTO(struct afs_call *call, const struct afs_fid *fid,
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
const struct qstr *name, const struct qstr *name2),
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TP_ARGS(call, fid, name, name2),
TP_STRUCT__entry(
__field(unsigned int, call )
__field(enum afs_fs_operation, op )
__field_struct(struct afs_fid, fid )
__array(char, name, 24 )
__array(char, name2, 24 )
),
TP_fast_assign(
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
unsigned int __len = min_t(unsigned int, name->len, 23);
unsigned int __len2 = min_t(unsigned int, name2->len, 23);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__entry->call = call->debug_id;
__entry->op = call->operation_ID;
if (fid) {
__entry->fid = *fid;
} else {
__entry->fid.vid = 0;
__entry->fid.vnode = 0;
__entry->fid.unique = 0;
}
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
memcpy(__entry->name, name->name, __len);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__entry->name[__len] = 0;
afs: Build an abstraction around an "operation" concept Turn the afs_operation struct into the main way that most fileserver operations are managed. Various things are added to the struct, including the following: (1) All the parameters and results of the relevant operations are moved into it, removing corresponding fields from the afs_call struct. afs_call gets a pointer to the op. (2) The target volume is made the main focus of the operation, rather than the target vnode(s), and a bunch of op->vnode->volume are made op->volume instead. (3) Two vnode records are defined (op->file[]) for the vnode(s) involved in most operations. The vnode record (struct afs_vnode_param) contains: - The vnode pointer. - The fid of the vnode to be included in the parameters or that was returned in the reply (eg. FS.MakeDir). - The status and callback information that may be returned in the reply about the vnode. - Callback break and data version tracking for detecting simultaneous third-parth changes. (4) Pointers to dentries to be updated with new inodes. (5) An operations table pointer. The table includes pointers to functions for issuing AFS and YFS-variant RPCs, handling the success and abort of an operation and handling post-I/O-lock local editing of a directory. To make this work, the following function restructuring is made: (A) The rotation loop that issues calls to fileservers that can be found in each function that wants to issue an RPC (such as afs_mkdir()) is extracted out into common code, in a new file called fs_operation.c. (B) The rotation loops, such as the one in afs_mkdir(), are replaced with a much smaller piece of code that allocates an operation, sets the parameters and then calls out to the common code to do the actual work. (C) The code for handling the success and failure of an operation are moved into operation functions (as (5) above) and these are called from the core code at appropriate times. (D) The pseudo inode getting stuff used by the dynamic root code is moved over into dynroot.c. (E) struct afs_iget_data is absorbed into the operation struct and afs_iget() expects to be given an op pointer and a vnode record. (F) Point (E) doesn't work for the root dir of a volume, but we know the FID in advance (it's always vnode 1, unique 1), so a separate inode getter, afs_root_iget(), is provided to special-case that. (G) The inode status init/update functions now also take an op and a vnode record. (H) The RPC marshalling functions now, for the most part, just take an afs_operation struct as their only argument. All the data they need is held there. The result delivery functions write their answers there as well. (I) The call is attached to the operation and then the operation core does the waiting. And then the new operation code is, for the moment, made to just initialise the operation, get the appropriate vnode I/O locks and do the same rotation loop as before. This lays the foundation for the following changes in the future: (*) Overhauling the rotation (again). (*) Support for asynchronous I/O, where the fileserver rotation must be done asynchronously also. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-11 03:51:51 +08:00
memcpy(__entry->name2, name2->name, __len2);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__entry->name2[__len2] = 0;
),
TP_printk("c=%08x %06llx:%06llx:%06x %s \"%s\" \"%s\"",
__entry->call,
__entry->fid.vid,
__entry->fid.vnode,
__entry->fid.unique,
__print_symbolic(__entry->op, afs_fs_operations),
__entry->name,
__entry->name2)
);
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
TRACE_EVENT(afs_make_vl_call,
TP_PROTO(struct afs_call *call),
TP_ARGS(call),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__field(enum afs_vl_operation, op )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->op = call->operation_ID;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk("c=%08x %s",
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->call,
__print_symbolic(__entry->op, afs_vl_operations))
);
TRACE_EVENT(afs_call_done,
TP_PROTO(struct afs_call *call),
TP_ARGS(call),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__field(struct rxrpc_call *, rx_call )
__field(int, ret )
__field(u32, abort_code )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->rx_call = call->rxcall;
__entry->ret = call->error;
__entry->abort_code = call->abort_code;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk(" c=%08x ret=%d ab=%d [%p]",
afs: Trace the initiation and completion of client calls Add tracepoints to trace the initiation and completion of client calls within the kafs filesystem. The afs_make_vl_call tracepoint watches calls to the volume location database server. The afs_make_fs_call tracepoint watches calls to the file server. The afs_call_done tracepoint watches for call completion. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->call,
__entry->ret,
__entry->abort_code,
__entry->rx_call)
);
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
TRACE_EVENT(afs_send_pages,
TP_PROTO(struct afs_call *call, struct msghdr *msg,
pgoff_t first, pgoff_t last, unsigned int offset),
TP_ARGS(call, msg, first, last, offset),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__field(pgoff_t, first )
__field(pgoff_t, last )
__field(unsigned int, nr )
__field(unsigned int, bytes )
__field(unsigned int, offset )
__field(unsigned int, flags )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->first = first;
__entry->last = last;
__entry->nr = msg->msg_iter.nr_segs;
__entry->bytes = msg->msg_iter.count;
__entry->offset = offset;
__entry->flags = msg->msg_flags;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk(" c=%08x %lx-%lx-%lx b=%x o=%x f=%x",
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->call,
__entry->first, __entry->first + __entry->nr - 1, __entry->last,
__entry->bytes, __entry->offset,
__entry->flags)
);
TRACE_EVENT(afs_sent_pages,
TP_PROTO(struct afs_call *call, pgoff_t first, pgoff_t last,
pgoff_t cursor, int ret),
TP_ARGS(call, first, last, cursor, ret),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__field(pgoff_t, first )
__field(pgoff_t, last )
__field(pgoff_t, cursor )
__field(int, ret )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->first = first;
__entry->last = last;
__entry->cursor = cursor;
__entry->ret = ret;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk(" c=%08x %lx-%lx c=%lx r=%d",
afs: Trace the sending of pages Add a pair of tracepoints to log the sending of pages for an FS.StoreData or FS.StoreData64 operation. Tracepoint afs_send_pages notes each set of pages added to the operation. There may be several of these per operation as we get up at most 8 contiguous pages in one go because the bvec we're using is on the stack. Tracepoint afs_sent_pages notes the end of adding data from a whole run of pages to the operation and the completion of the request phase. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:51 +08:00
__entry->call,
__entry->first, __entry->last,
__entry->cursor, __entry->ret)
);
afs: Fix directory read/modify race Because parsing of the directory wasn't being done under any sort of lock, the pages holding the directory content can get invalidated whilst the parsing is ongoing. Further, the directory page check function gets called outside of the page lock, so if the page gets cleared or updated, this may return reports of bad magic numbers in the directory page. Also, the directory may change size whilst checking and parsing are ongoing, so more care needs to be taken here. Fix this by: (1) Perform the page check from the page filling function before we set PageUptodate and drop the page lock. (2) Check for the file having shrunk and the page having been abandoned before checking the page contents. (3) Lock the page whilst parsing it for the directory iterator. Whilst we're at it, add a tracepoint to report check failure. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:52 +08:00
TRACE_EVENT(afs_dir_check_failed,
TP_PROTO(struct afs_vnode *vnode, loff_t off, loff_t i_size),
TP_ARGS(vnode, off, i_size),
TP_STRUCT__entry(
__field(struct afs_vnode *, vnode )
__field(loff_t, off )
__field(loff_t, i_size )
),
TP_fast_assign(
__entry->vnode = vnode;
__entry->off = off;
__entry->i_size = i_size;
),
TP_printk("vn=%p %llx/%llx",
__entry->vnode, __entry->off, __entry->i_size)
);
afs: Trace page dirty/clean Add a trace event that logs the dirtying and cleaning of pages attached to AFS inodes. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:53 +08:00
/*
* We use page->private to hold the amount of the page that we've written to,
* splitting the field into two parts. However, we need to represent a range
* 0...PAGE_SIZE inclusive, so we can't support 64K pages on a 32-bit system.
*/
#if PAGE_SIZE > 32768
#define AFS_PRIV_MAX 0xffffffff
#define AFS_PRIV_SHIFT 32
#else
#define AFS_PRIV_MAX 0xffff
#define AFS_PRIV_SHIFT 16
#endif
TRACE_EVENT(afs_page_dirty,
TP_PROTO(struct afs_vnode *vnode, const char *where,
pgoff_t page, unsigned long priv),
TP_ARGS(vnode, where, page, priv),
TP_STRUCT__entry(
__field(struct afs_vnode *, vnode )
__field(const char *, where )
__field(pgoff_t, page )
__field(unsigned long, priv )
),
TP_fast_assign(
__entry->vnode = vnode;
__entry->where = where;
__entry->page = page;
__entry->priv = priv;
),
TP_printk("vn=%p %lx %s %lu-%lu",
__entry->vnode, __entry->page, __entry->where,
__entry->priv & AFS_PRIV_MAX,
__entry->priv >> AFS_PRIV_SHIFT)
);
afs: Protect call->state changes against signals Protect call->state changes against the call being prematurely terminated due to a signal. What can happen is that a signal causes afs_wait_for_call_to_complete() to abort an afs_call because it's not yet complete whilst afs_deliver_to_call() is delivering data to that call. If the data delivery causes the state to change, this may overwrite the state of the afs_call, making it not-yet-complete again - but no further notifications will be forthcoming from AF_RXRPC as the rxrpc call has been aborted and completed, so kAFS will just hang in various places waiting for that call or on page bits that need clearing by that call. A tracepoint to monitor call state changes is also provided. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:53 +08:00
TRACE_EVENT(afs_call_state,
TP_PROTO(struct afs_call *call,
enum afs_call_state from,
enum afs_call_state to,
int ret, u32 remote_abort),
TP_ARGS(call, from, to, ret, remote_abort),
TP_STRUCT__entry(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__field(unsigned int, call )
afs: Protect call->state changes against signals Protect call->state changes against the call being prematurely terminated due to a signal. What can happen is that a signal causes afs_wait_for_call_to_complete() to abort an afs_call because it's not yet complete whilst afs_deliver_to_call() is delivering data to that call. If the data delivery causes the state to change, this may overwrite the state of the afs_call, making it not-yet-complete again - but no further notifications will be forthcoming from AF_RXRPC as the rxrpc call has been aborted and completed, so kAFS will just hang in various places waiting for that call or on page bits that need clearing by that call. A tracepoint to monitor call state changes is also provided. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:53 +08:00
__field(enum afs_call_state, from )
__field(enum afs_call_state, to )
__field(int, ret )
__field(u32, abort )
),
TP_fast_assign(
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
__entry->call = call->debug_id;
afs: Protect call->state changes against signals Protect call->state changes against the call being prematurely terminated due to a signal. What can happen is that a signal causes afs_wait_for_call_to_complete() to abort an afs_call because it's not yet complete whilst afs_deliver_to_call() is delivering data to that call. If the data delivery causes the state to change, this may overwrite the state of the afs_call, making it not-yet-complete again - but no further notifications will be forthcoming from AF_RXRPC as the rxrpc call has been aborted and completed, so kAFS will just hang in various places waiting for that call or on page bits that need clearing by that call. A tracepoint to monitor call state changes is also provided. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:53 +08:00
__entry->from = from;
__entry->to = to;
__entry->ret = ret;
__entry->abort = remote_abort;
),
rxrpc, afs: Use debug_ids rather than pointers in traces In rxrpc and afs, use the debug_ids that are monotonically allocated to various objects as they're allocated rather than pointers as kernel pointers are now hashed making them less useful. Further, the debug ids aren't reused anywhere nearly as quickly. In addition, allow kernel services that use rxrpc, such as afs, to take numbers from the rxrpc counter, assign them to their own call struct and pass them in to rxrpc for both client and service calls so that the trace lines for each will have the same ID tag. Signed-off-by: David Howells <dhowells@redhat.com>
2018-03-28 06:03:00 +08:00
TP_printk("c=%08x %u->%u r=%d ab=%d",
afs: Protect call->state changes against signals Protect call->state changes against the call being prematurely terminated due to a signal. What can happen is that a signal causes afs_wait_for_call_to_complete() to abort an afs_call because it's not yet complete whilst afs_deliver_to_call() is delivering data to that call. If the data delivery causes the state to change, this may overwrite the state of the afs_call, making it not-yet-complete again - but no further notifications will be forthcoming from AF_RXRPC as the rxrpc call has been aborted and completed, so kAFS will just hang in various places waiting for that call or on page bits that need clearing by that call. A tracepoint to monitor call state changes is also provided. Signed-off-by: David Howells <dhowells@redhat.com>
2017-11-02 23:27:53 +08:00
__entry->call,
__entry->from, __entry->to,
__entry->ret, __entry->abort)
);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TRACE_EVENT(afs_lookup,
TP_PROTO(struct afs_vnode *dvnode, const struct qstr *name,
afs: Fix use-after-loss-of-ref afs_lookup() has a tracepoint to indicate the outcome of d_splice_alias(), passing it the inode to retrieve the fid from. However, the function gave up its ref on that inode when it called d_splice_alias(), which may have failed and dropped the inode. Fix this by caching the fid. Fixes: 80548b03991f ("afs: Add more tracepoints") Reported-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-15 00:16:25 +08:00
struct afs_fid *fid),
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
afs: Fix use-after-loss-of-ref afs_lookup() has a tracepoint to indicate the outcome of d_splice_alias(), passing it the inode to retrieve the fid from. However, the function gave up its ref on that inode when it called d_splice_alias(), which may have failed and dropped the inode. Fix this by caching the fid. Fixes: 80548b03991f ("afs: Add more tracepoints") Reported-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-15 00:16:25 +08:00
TP_ARGS(dvnode, name, fid),
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TP_STRUCT__entry(
__field_struct(struct afs_fid, dfid )
__field_struct(struct afs_fid, fid )
__array(char, name, 24 )
),
TP_fast_assign(
int __len = min_t(int, name->len, 23);
__entry->dfid = dvnode->fid;
afs: Fix use-after-loss-of-ref afs_lookup() has a tracepoint to indicate the outcome of d_splice_alias(), passing it the inode to retrieve the fid from. However, the function gave up its ref on that inode when it called d_splice_alias(), which may have failed and dropped the inode. Fix this by caching the fid. Fixes: 80548b03991f ("afs: Add more tracepoints") Reported-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-15 00:16:25 +08:00
__entry->fid = *fid;
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
memcpy(__entry->name, name->name, __len);
__entry->name[__len] = 0;
),
TP_printk("d=%llx:%llx:%x \"%s\" f=%llx:%x",
__entry->dfid.vid, __entry->dfid.vnode, __entry->dfid.unique,
__entry->name,
__entry->fid.vnode, __entry->fid.unique)
);
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
TRACE_EVENT(afs_edit_dir,
TP_PROTO(struct afs_vnode *dvnode,
enum afs_edit_dir_reason why,
enum afs_edit_dir_op op,
unsigned int block,
unsigned int slot,
unsigned int f_vnode,
unsigned int f_unique,
const char *name),
TP_ARGS(dvnode, why, op, block, slot, f_vnode, f_unique, name),
TP_STRUCT__entry(
__field(unsigned int, vnode )
__field(unsigned int, unique )
__field(enum afs_edit_dir_reason, why )
__field(enum afs_edit_dir_op, op )
__field(unsigned int, block )
__field(unsigned short, slot )
__field(unsigned int, f_vnode )
__field(unsigned int, f_unique )
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__array(char, name, 24 )
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
),
TP_fast_assign(
int __len = strlen(name);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
__len = min(__len, 23);
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
__entry->vnode = dvnode->fid.vnode;
__entry->unique = dvnode->fid.unique;
__entry->why = why;
__entry->op = op;
__entry->block = block;
__entry->slot = slot;
__entry->f_vnode = f_vnode;
__entry->f_unique = f_unique;
memcpy(__entry->name, name, __len);
__entry->name[__len] = 0;
),
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TP_printk("d=%x:%x %s %s %u[%u] f=%x:%x \"%s\"",
afs: Locally edit directory data for mkdir/create/unlink/... Locally edit the contents of an AFS directory upon a successful inode operation that modifies that directory (such as mkdir, create and unlink) so that we can avoid the current practice of re-downloading the directory after each change. This is viable provided that the directory version number we get back from the modifying RPC op is exactly incremented by 1 from what we had previously. The data in the directory contents is in a defined format that we have to parse locally to perform lookups and readdir, so modifying isn't a problem. If the edit fails, we just clear the VALID flag on the directory and it will be reloaded next time it is needed. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
__entry->vnode, __entry->unique,
__print_symbolic(__entry->why, afs_edit_dir_reasons),
__print_symbolic(__entry->op, afs_edit_dir_ops),
__entry->block, __entry->slot,
__entry->f_vnode, __entry->f_unique,
__entry->name)
);
afs: Trace protocol errors Trace protocol errors detected in afs. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
TRACE_EVENT(afs_protocol_error,
afs: Remove the error argument from afs_protocol_error() Remove the error argument from afs_protocol_error() as it's always -EBADMSG. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-08 23:49:08 +08:00
TP_PROTO(struct afs_call *call, enum afs_eproto_cause cause),
afs: Trace protocol errors Trace protocol errors detected in afs. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
afs: Remove the error argument from afs_protocol_error() Remove the error argument from afs_protocol_error() as it's always -EBADMSG. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-08 23:49:08 +08:00
TP_ARGS(call, cause),
afs: Trace protocol errors Trace protocol errors detected in afs. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
TP_STRUCT__entry(
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
__field(unsigned int, call )
__field(enum afs_eproto_cause, cause )
afs: Trace protocol errors Trace protocol errors detected in afs. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
),
TP_fast_assign(
__entry->call = call ? call->debug_id : 0;
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
__entry->cause = cause;
afs: Trace protocol errors Trace protocol errors detected in afs. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
),
afs: Remove the error argument from afs_protocol_error() Remove the error argument from afs_protocol_error() as it's always -EBADMSG. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-08 23:49:08 +08:00
TP_printk("c=%08x %s",
__entry->call,
afs: Better tracing of protocol errors Include the site of detection of AFS protocol errors in trace lines to better be able to determine what went wrong. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:56 +08:00
__print_symbolic(__entry->cause, afs_eproto_causes))
afs: Trace protocol errors Trace protocol errors detected in afs. Signed-off-by: David Howells <dhowells@redhat.com>
2018-04-06 21:17:25 +08:00
);
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
TRACE_EVENT(afs_io_error,
TP_PROTO(unsigned int call, int error, enum afs_io_error where),
TP_ARGS(call, error, where),
TP_STRUCT__entry(
__field(unsigned int, call )
__field(int, error )
__field(enum afs_io_error, where )
),
TP_fast_assign(
__entry->call = call;
__entry->error = error;
__entry->where = where;
),
TP_printk("c=%08x r=%d %s",
__entry->call, __entry->error,
__print_symbolic(__entry->where, afs_io_errors))
);
TRACE_EVENT(afs_file_error,
TP_PROTO(struct afs_vnode *vnode, int error, enum afs_file_error where),
TP_ARGS(vnode, error, where),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
__field(int, error )
__field(enum afs_file_error, where )
),
TP_fast_assign(
__entry->fid = vnode->fid;
__entry->error = error;
__entry->where = where;
),
afs: Increase to 64-bit volume ID and 96-bit vnode ID for YFS Increase the sizes of the volume ID to 64 bits and the vnode ID (inode number equivalent) to 96 bits to allow the support of YFS. This requires the iget comparator to check the vnode->fid rather than i_ino and i_generation as i_ino is not sufficiently capacious. It also requires this data to be placed into the vnode cache key for fscache. For the moment, just discard the top 32 bits of the vnode ID when returning it though stat. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
TP_printk("%llx:%llx:%x r=%d %s",
afs: Add a couple of tracepoints to log I/O errors Add a couple of tracepoints to log the production of I/O errors within the AFS filesystem. Signed-off-by: David Howells <dhowells@redhat.com>
2018-10-20 07:57:57 +08:00
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique,
__entry->error,
__print_symbolic(__entry->where, afs_file_errors))
);
afs: Add a tracepoint to record callbacks from unlisted servers Add a tracepoint to record callbacks from servers for which we don't have a record. Signed-off-by: David Howells <dhowells@redhat.com>
2018-05-12 05:59:42 +08:00
TRACE_EVENT(afs_cm_no_server,
TP_PROTO(struct afs_call *call, struct sockaddr_rxrpc *srx),
TP_ARGS(call, srx),
TP_STRUCT__entry(
__field(unsigned int, call )
__field(unsigned int, op_id )
__field_struct(struct sockaddr_rxrpc, srx )
),
TP_fast_assign(
__entry->call = call->debug_id;
__entry->op_id = call->operation_ID;
memcpy(&__entry->srx, srx, sizeof(__entry->srx));
),
TP_printk("c=%08x op=%u %pISpc",
__entry->call, __entry->op_id, &__entry->srx.transport)
);
TRACE_EVENT(afs_cm_no_server_u,
TP_PROTO(struct afs_call *call, const uuid_t *uuid),
TP_ARGS(call, uuid),
TP_STRUCT__entry(
__field(unsigned int, call )
__field(unsigned int, op_id )
__field_struct(uuid_t, uuid )
),
TP_fast_assign(
__entry->call = call->debug_id;
__entry->op_id = call->operation_ID;
memcpy(&__entry->uuid, uuid, sizeof(__entry->uuid));
),
TP_printk("c=%08x op=%u %pU",
__entry->call, __entry->op_id, &__entry->uuid)
);
afs: Add file locking tracepoints Add two tracepoints for monitoring AFS file locking. Firstly, add one that follows the operational part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_op/enable And add a second that more follows the event-driven part: echo 1 >/sys/kernel/debug/tracing/events/afs/afs_flock_ev/enable Individual file_lock structs seen by afs are tagged with debugging IDs that are displayed in the trace log to make it easier to see what's going on, especially as setting the first lock always seems to involve copying the file_lock twice. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:50 +08:00
TRACE_EVENT(afs_flock_ev,
TP_PROTO(struct afs_vnode *vnode, struct file_lock *fl,
enum afs_flock_event event, int error),
TP_ARGS(vnode, fl, event, error),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
__field(enum afs_flock_event, event )
__field(enum afs_lock_state, state )
__field(int, error )
__field(unsigned int, debug_id )
),
TP_fast_assign(
__entry->fid = vnode->fid;
__entry->event = event;
__entry->state = vnode->lock_state;
__entry->error = error;
__entry->debug_id = fl ? fl->fl_u.afs.debug_id : 0;
),
TP_printk("%llx:%llx:%x %04x %s s=%s e=%d",
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique,
__entry->debug_id,
__print_symbolic(__entry->event, afs_flock_events),
__print_symbolic(__entry->state, afs_flock_states),
__entry->error)
);
TRACE_EVENT(afs_flock_op,
TP_PROTO(struct afs_vnode *vnode, struct file_lock *fl,
enum afs_flock_operation op),
TP_ARGS(vnode, fl, op),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
__field(loff_t, from )
__field(loff_t, len )
__field(enum afs_flock_operation, op )
__field(unsigned char, type )
__field(unsigned int, flags )
__field(unsigned int, debug_id )
),
TP_fast_assign(
__entry->fid = vnode->fid;
__entry->from = fl->fl_start;
__entry->len = fl->fl_end - fl->fl_start + 1;
__entry->op = op;
__entry->type = fl->fl_type;
__entry->flags = fl->fl_flags;
__entry->debug_id = fl->fl_u.afs.debug_id;
),
TP_printk("%llx:%llx:%x %04x %s t=%s R=%llx/%llx f=%x",
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique,
__entry->debug_id,
__print_symbolic(__entry->op, afs_flock_operations),
__print_symbolic(__entry->type, afs_flock_types),
__entry->from, __entry->len, __entry->flags)
);
afs: Add directory reload tracepoint Add a tracepoint (afs_reload_dir) to indicate when a directory is being reloaded. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TRACE_EVENT(afs_reload_dir,
TP_PROTO(struct afs_vnode *vnode),
TP_ARGS(vnode),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
),
TP_fast_assign(
__entry->fid = vnode->fid;
),
TP_printk("%llx:%llx:%x",
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique)
);
afs: Implement sillyrename for unlink and rename Implement sillyrename for AFS unlink and rename, using the NFS variant implementation as a basis. Note that the asynchronous file locking extender/releaser has to be notified with a state change to stop it complaining if there's a race between that and the actual file deletion. A tracepoint, afs_silly_rename, is also added to note the silly rename and the cleanup. The afs_edit_dir tracepoint is given some extra reason indicators and the afs_flock_ev tracepoint is given a silly-delete file lock cancellation indicator. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TRACE_EVENT(afs_silly_rename,
TP_PROTO(struct afs_vnode *vnode, bool done),
TP_ARGS(vnode, done),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
__field(bool, done )
),
TP_fast_assign(
__entry->fid = vnode->fid;
__entry->done = done;
),
TP_printk("%llx:%llx:%x done=%u",
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique,
__entry->done)
);
afs: Add more tracepoints Add four more tracepoints: (1) afs_make_fs_call1 - Split from afs_make_fs_call but takes a filename to log also. (2) afs_make_fs_call2 - Like the above but takes two filenames to log. (3) afs_lookup - Log the result of doing a successful lookup, including a negative result (fid 0:0). (4) afs_get_tree - Log the set up of a volume for mounting. It also extends the name buffer on the afs_edit_dir tracepoint to 24 chars and puts quotes around the filename in the text representation. Signed-off-by: David Howells <dhowells@redhat.com>
2019-04-25 21:26:51 +08:00
TRACE_EVENT(afs_get_tree,
TP_PROTO(struct afs_cell *cell, struct afs_volume *volume),
TP_ARGS(cell, volume),
TP_STRUCT__entry(
__field(u64, vid )
__array(char, cell, 24 )
__array(char, volume, 24 )
),
TP_fast_assign(
int __len;
__entry->vid = volume->vid;
__len = min_t(int, cell->name_len, 23);
memcpy(__entry->cell, cell->name, __len);
__entry->cell[__len] = 0;
__len = min_t(int, volume->name_len, 23);
memcpy(__entry->volume, volume->name, __len);
__entry->volume[__len] = 0;
),
TP_printk("--- MOUNT %s:%s %llx",
__entry->cell, __entry->volume, __entry->vid)
);
afs: Add some callback management tracepoints Add a couple of tracepoints to track callback management: (1) afs_cb_miss - Logs when we were unable to apply a callback, either due to the inode being discarded or due to a competing thread applying a callback first. (2) afs_cb_break - Logs when we attempted to clear the noted callback promise, either due to the server explicitly breaking the callback, the callback promise lapsing or a local event obsoleting it. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:16 +08:00
TRACE_EVENT(afs_cb_break,
TP_PROTO(struct afs_fid *fid, unsigned int cb_break,
enum afs_cb_break_reason reason, bool skipped),
TP_ARGS(fid, cb_break, reason, skipped),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
__field(unsigned int, cb_break )
__field(enum afs_cb_break_reason, reason )
__field(bool, skipped )
),
TP_fast_assign(
__entry->fid = *fid;
__entry->cb_break = cb_break;
__entry->reason = reason;
__entry->skipped = skipped;
),
TP_printk("%llx:%llx:%x b=%x s=%u %s",
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique,
__entry->cb_break,
__entry->skipped,
__print_symbolic(__entry->reason, afs_cb_break_reasons))
);
TRACE_EVENT(afs_cb_miss,
TP_PROTO(struct afs_fid *fid, enum afs_cb_break_reason reason),
TP_ARGS(fid, reason),
TP_STRUCT__entry(
__field_struct(struct afs_fid, fid )
__field(enum afs_cb_break_reason, reason )
),
TP_fast_assign(
__entry->fid = *fid;
__entry->reason = reason;
),
TP_printk(" %llx:%llx:%x %s",
__entry->fid.vid, __entry->fid.vnode, __entry->fid.unique,
__print_symbolic(__entry->reason, afs_cb_break_reasons))
);
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
TRACE_EVENT(afs_server,
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
TP_PROTO(struct afs_server *server, int ref, int active,
enum afs_server_trace reason),
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
TP_ARGS(server, ref, active, reason),
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
TP_STRUCT__entry(
__field(unsigned int, server )
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
__field(int, ref )
__field(int, active )
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
__field(int, reason )
),
TP_fast_assign(
__entry->server = server->debug_id;
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
__entry->ref = ref;
__entry->active = active;
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
__entry->reason = reason;
),
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
TP_printk("s=%08x %s u=%d a=%d",
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
__entry->server,
__print_symbolic(__entry->reason, afs_server_traces),
afs: Split the usage count on struct afs_server Split the usage count on the afs_server struct to have an active count that registers who's actually using it separately from the reference count on the object. This allows a future patch to dispatch polling probes without advancing the "unuse" time into the future each time we emit a probe, which would otherwise prevent unused server records from expiring. Included in this: (1) The latter part of afs_destroy_server() in which the RCU destruction of afs_server objects is invoked and the outstanding server count is decremented is split out into __afs_put_server(). (2) afs_put_server() now calls __afs_put_server() rather then setting the management timer. (3) The calls begun by afs_fs_give_up_all_callbacks() and afs_fs_get_capabilities() can now take a ref on the server record, so afs_destroy_server() can just drop its ref and needn't wait for the completion of these calls. They'll put the ref when they're done. (4) Because of (3), afs_fs_probe_done() no longer needs to wake up afs_destroy_server() with server->probe_outstanding. (5) afs_gc_servers can be simplified. It only needs to check if server->active is 0 rather than playing games with the refcount. (6) afs_manage_servers() can propose a server for gc if usage == 0 rather than if ref == 1. The gc is effected by (5). Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-18 00:31:26 +08:00
__entry->ref,
__entry->active)
afs: Trace afs_server usage Add a tracepoint (afs_server) to track the afs_server object usage count. Signed-off-by: David Howells <dhowells@redhat.com>
2019-06-21 01:12:17 +08:00
);
afs: Add a tracepoint to track the lifetime of the afs_volume struct Add a tracepoint to track the lifetime of the afs_volume struct. Signed-off-by: David Howells <dhowells@redhat.com>
2020-04-30 00:02:04 +08:00
TRACE_EVENT(afs_volume,
TP_PROTO(afs_volid_t vid, int ref, enum afs_volume_trace reason),
TP_ARGS(vid, ref, reason),
TP_STRUCT__entry(
__field(afs_volid_t, vid )
__field(int, ref )
__field(enum afs_volume_trace, reason )
),
TP_fast_assign(
__entry->vid = vid;
__entry->ref = ref;
__entry->reason = reason;
),
TP_printk("V=%llx %s u=%d",
__entry->vid,
__print_symbolic(__entry->reason, afs_volume_traces),
__entry->ref)
);
afs: Add some tracepoints Add three tracepoints to the AFS filesystem: (1) The afs_recv_data tracepoint logs data segments that are extracted from the data received from the peer through afs_extract_data(). (2) The afs_notify_call tracepoint logs notification from AF_RXRPC of data coming in to an asynchronous call. (3) The afs_cb_call tracepoint logs incoming calls that have had their operation ID extracted and mapped into a supported cache manager service call. To make (3) work, the name strings in the afs_call_type struct objects have to be annotated with __tracepoint_string. This is done with the CM_NAME() macro. Further, the AFS call state enum needs a name so that it can be used to declare parameter types. Signed-off-by: David Howells <dhowells@redhat.com>
2017-01-05 18:38:34 +08:00
#endif /* _TRACE_AFS_H */
/* This part must be outside protection */
#include <trace/define_trace.h>