• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1                   Glock internal locking rules
2                  ------------------------------
3
4This documents the basic principles of the glock state machine
5internals. Each glock (struct gfs2_glock in fs/gfs2/incore.h)
6has two main (internal) locks:
7
8 1. A spinlock (gl_lockref.lock) which protects the internal state such
9    as gl_state, gl_target and the list of holders (gl_holders)
10 2. A non-blocking bit lock, GLF_LOCK, which is used to prevent other
11    threads from making calls to the DLM, etc. at the same time. If a
12    thread takes this lock, it must then call run_queue (usually via the
13    workqueue) when it releases it in order to ensure any pending tasks
14    are completed.
15
16The gl_holders list contains all the queued lock requests (not
17just the holders) associated with the glock. If there are any
18held locks, then they will be contiguous entries at the head
19of the list. Locks are granted in strictly the order that they
20are queued, except for those marked LM_FLAG_PRIORITY which are
21used only during recovery, and even then only for journal locks.
22
23There are three lock states that users of the glock layer can request,
24namely shared (SH), deferred (DF) and exclusive (EX). Those translate
25to the following DLM lock modes:
26
27Glock mode    | DLM lock mode
28------------------------------
29    UN        |    IV/NL  Unlocked (no DLM lock associated with glock) or NL
30    SH        |    PR     (Protected read)
31    DF        |    CW     (Concurrent write)
32    EX        |    EX     (Exclusive)
33
34Thus DF is basically a shared mode which is incompatible with the "normal"
35shared lock mode, SH. In GFS2 the DF mode is used exclusively for direct I/O
36operations. The glocks are basically a lock plus some routines which deal
37with cache management. The following rules apply for the cache:
38
39Glock mode   |  Cache data | Cache Metadata | Dirty Data | Dirty Metadata
40--------------------------------------------------------------------------
41    UN       |     No      |       No       |     No     |      No
42    SH       |     Yes     |       Yes      |     No     |      No
43    DF       |     No      |       Yes      |     No     |      No
44    EX       |     Yes     |       Yes      |     Yes    |      Yes
45
46These rules are implemented using the various glock operations which
47are defined for each type of glock. Not all types of glocks use
48all the modes. Only inode glocks use the DF mode for example.
49
50Table of glock operations and per type constants:
51
52Field            | Purpose
53----------------------------------------------------------------------------
54go_xmote_th      | Called before remote state change (e.g. to sync dirty data)
55go_xmote_bh      | Called after remote state change (e.g. to refill cache)
56go_inval         | Called if remote state change requires invalidating the cache
57go_demote_ok     | Returns boolean value of whether its ok to demote a glock
58                 | (e.g. checks timeout, and that there is no cached data)
59go_lock          | Called for the first local holder of a lock
60go_unlock        | Called on the final local unlock of a lock
61go_dump          | Called to print content of object for debugfs file, or on
62                 | error to dump glock to the log.
63go_type          | The type of the glock, LM_TYPE_.....
64go_callback	 | Called if the DLM sends a callback to drop this lock
65go_flags	 | GLOF_ASPACE is set, if the glock has an address space
66                 | associated with it
67
68The minimum hold time for each lock is the time after a remote lock
69grant for which we ignore remote demote requests. This is in order to
70prevent a situation where locks are being bounced around the cluster
71from node to node with none of the nodes making any progress. This
72tends to show up most with shared mmaped files which are being written
73to by multiple nodes. By delaying the demotion in response to a
74remote callback, that gives the userspace program time to make
75some progress before the pages are unmapped.
76
77There is a plan to try and remove the go_lock and go_unlock callbacks
78if possible, in order to try and speed up the fast path though the locking.
79Also, eventually we hope to make the glock "EX" mode locally shared
80such that any local locking will be done with the i_mutex as required
81rather than via the glock.
82
83Locking rules for glock operations:
84
85Operation     |  GLF_LOCK bit lock held |  gl_lockref.lock spinlock held
86-------------------------------------------------------------------------
87go_xmote_th   |       Yes               |       No
88go_xmote_bh   |       Yes               |       No
89go_inval      |       Yes               |       No
90go_demote_ok  |       Sometimes         |       Yes
91go_lock       |       Yes               |       No
92go_unlock     |       Yes               |       No
93go_dump       |       Sometimes         |       Yes
94go_callback   |       Sometimes (N/A)   |       Yes
95
96N.B. Operations must not drop either the bit lock or the spinlock
97if its held on entry. go_dump and do_demote_ok must never block.
98Note that go_dump will only be called if the glock's state
99indicates that it is caching uptodate data.
100
101Glock locking order within GFS2:
102
103 1. i_rwsem (if required)
104 2. Rename glock (for rename only)
105 3. Inode glock(s)
106    (Parents before children, inodes at "same level" with same parent in
107     lock number order)
108 4. Rgrp glock(s) (for (de)allocation operations)
109 5. Transaction glock (via gfs2_trans_begin) for non-read operations
110 6. i_rw_mutex (if required)
111 7. Page lock  (always last, very important!)
112
113There are two glocks per inode. One deals with access to the inode
114itself (locking order as above), and the other, known as the iopen
115glock is used in conjunction with the i_nlink field in the inode to
116determine the lifetime of the inode in question. Locking of inodes
117is on a per-inode basis. Locking of rgrps is on a per rgrp basis.
118In general we prefer to lock local locks prior to cluster locks.
119
120                            Glock Statistics
121                           ------------------
122
123The stats are divided into two sets: those relating to the
124super block and those relating to an individual glock. The
125super block stats are done on a per cpu basis in order to
126try and reduce the overhead of gathering them. They are also
127further divided by glock type. All timings are in nanoseconds.
128
129In the case of both the super block and glock statistics,
130the same information is gathered in each case. The super
131block timing statistics are used to provide default values for
132the glock timing statistics, so that newly created glocks
133should have, as far as possible, a sensible starting point.
134The per-glock counters are initialised to zero when the
135glock is created. The per-glock statistics are lost when
136the glock is ejected from memory.
137
138The statistics are divided into three pairs of mean and
139variance, plus two counters. The mean/variance pairs are
140smoothed exponential estimates and the algorithm used is
141one which will be very familiar to those used to calculation
142of round trip times in network code. See "TCP/IP Illustrated,
143Volume 1", W. Richard Stevens, sect 21.3, "Round-Trip Time Measurement",
144p. 299 and onwards. Also, Volume 2, Sect. 25.10, p. 838 and onwards.
145Unlike the TCP/IP Illustrated case, the mean and variance are
146not scaled, but are in units of integer nanoseconds.
147
148The three pairs of mean/variance measure the following
149things:
150
151 1. DLM lock time (non-blocking requests)
152 2. DLM lock time (blocking requests)
153 3. Inter-request time (again to the DLM)
154
155A non-blocking request is one which will complete right
156away, whatever the state of the DLM lock in question. That
157currently means any requests when (a) the current state of
158the lock is exclusive, i.e. a lock demotion (b) the requested
159state is either null or unlocked (again, a demotion) or (c) the
160"try lock" flag is set. A blocking request covers all the other
161lock requests.
162
163There are two counters. The first is there primarily to show
164how many lock requests have been made, and thus how much data
165has gone into the mean/variance calculations. The other counter
166is counting queuing of holders at the top layer of the glock
167code. Hopefully that number will be a lot larger than the number
168of dlm lock requests issued.
169
170So why gather these statistics? There are several reasons
171we'd like to get a better idea of these timings:
172
1731. To be able to better set the glock "min hold time"
1742. To spot performance issues more easily
1753. To improve the algorithm for selecting resource groups for
176allocation (to base it on lock wait time, rather than blindly
177using a "try lock")
178
179Due to the smoothing action of the updates, a step change in
180some input quantity being sampled will only fully be taken
181into account after 8 samples (or 4 for the variance) and this
182needs to be carefully considered when interpreting the
183results.
184
185Knowing both the time it takes a lock request to complete and
186the average time between lock requests for a glock means we
187can compute the total percentage of the time for which the
188node is able to use a glock vs. time that the rest of the
189cluster has its share. That will be very useful when setting
190the lock min hold time.
191
192Great care has been taken to ensure that we
193measure exactly the quantities that we want, as accurately
194as possible. There are always inaccuracies in any
195measuring system, but I hope this is as accurate as we
196can reasonably make it.
197
198Per sb stats can be found here:
199/sys/kernel/debug/gfs2/<fsname>/sbstats
200Per glock stats can be found here:
201/sys/kernel/debug/gfs2/<fsname>/glstats
202
203Assuming that debugfs is mounted on /sys/kernel/debug and also
204that <fsname> is replaced with the name of the gfs2 filesystem
205in question.
206
207The abbreviations used in the output as are follows:
208
209srtt     - Smoothed round trip time for non-blocking dlm requests
210srttvar  - Variance estimate for srtt
211srttb    - Smoothed round trip time for (potentially) blocking dlm requests
212srttvarb - Variance estimate for srttb
213sirt     - Smoothed inter-request time (for dlm requests)
214sirtvar  - Variance estimate for sirt
215dlm      - Number of dlm requests made (dcnt in glstats file)
216queue    - Number of glock requests queued (qcnt in glstats file)
217
218The sbstats file contains a set of these stats for each glock type (so 8 lines
219for each type) and for each cpu (one column per cpu). The glstats file contains
220a set of these stats for each glock in a similar format to the glocks file, but
221using the format mean/variance for each of the timing stats.
222
223The gfs2_glock_lock_time tracepoint prints out the current values of the stats
224for the glock in question, along with some addition information on each dlm
225reply that is received:
226
227status - The status of the dlm request
228flags  - The dlm request flags
229tdiff  - The time taken by this specific request
230(remaining fields as per above list)
231
232
233