1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 ======= 3 =======
4 SCSI EH 4 SCSI EH
5 ======= 5 =======
6 6
7 This document describes SCSI midlayer error ha 7 This document describes SCSI midlayer error handling infrastructure.
8 Please refer to Documentation/scsi/scsi_mid_lo 8 Please refer to Documentation/scsi/scsi_mid_low_api.rst for more
9 information regarding SCSI midlayer. 9 information regarding SCSI midlayer.
10 10
11 .. TABLE OF CONTENTS 11 .. TABLE OF CONTENTS
12 12
13 [1] How SCSI commands travel through the mi 13 [1] How SCSI commands travel through the midlayer and to EH
14 [1-1] struct scsi_cmnd 14 [1-1] struct scsi_cmnd
15 [1-2] How do scmd's get completed? 15 [1-2] How do scmd's get completed?
16 [1-2-1] Completing a scmd w/ scsi_done 16 [1-2-1] Completing a scmd w/ scsi_done
17 [1-2-2] Completing a scmd w/ timeout 17 [1-2-2] Completing a scmd w/ timeout
18 [1-3] How EH takes over 18 [1-3] How EH takes over
19 [2] How SCSI EH works 19 [2] How SCSI EH works
20 [2-1] EH through fine-grained callbacks 20 [2-1] EH through fine-grained callbacks
21 [2-1-1] Overview 21 [2-1-1] Overview
22 [2-1-2] Flow of scmds through EH 22 [2-1-2] Flow of scmds through EH
23 [2-1-3] Flow of control 23 [2-1-3] Flow of control
24 [2-2] EH through transportt->eh_strateg 24 [2-2] EH through transportt->eh_strategy_handler()
25 [2-2-1] Pre transportt->eh_strategy_ha 25 [2-2-1] Pre transportt->eh_strategy_handler() SCSI midlayer conditions
26 [2-2-2] Post transportt->eh_strategy_h 26 [2-2-2] Post transportt->eh_strategy_handler() SCSI midlayer conditions
27 [2-2-3] Things to consider 27 [2-2-3] Things to consider
28 28
29 29
30 1. How SCSI commands travel through the midlay 30 1. How SCSI commands travel through the midlayer and to EH
31 ============================================== 31 ==========================================================
32 32
33 1.1 struct scsi_cmnd 33 1.1 struct scsi_cmnd
34 -------------------- 34 --------------------
35 35
36 Each SCSI command is represented with struct s 36 Each SCSI command is represented with struct scsi_cmnd (== scmd). A
37 scmd has two list_head's to link itself into l 37 scmd has two list_head's to link itself into lists. The two are
38 scmd->list and scmd->eh_entry. The former is 38 scmd->list and scmd->eh_entry. The former is used for free list or
39 per-device allocated scmd list and not of much 39 per-device allocated scmd list and not of much interest to this EH
40 discussion. The latter is used for completion 40 discussion. The latter is used for completion and EH lists and unless
41 otherwise stated scmds are always linked using 41 otherwise stated scmds are always linked using scmd->eh_entry in this
42 discussion. 42 discussion.
43 43
44 44
45 1.2 How do scmd's get completed? 45 1.2 How do scmd's get completed?
46 -------------------------------- 46 --------------------------------
47 47
48 Once LLDD gets hold of a scmd, either the LLDD 48 Once LLDD gets hold of a scmd, either the LLDD will complete the
49 command by calling scsi_done callback passed f 49 command by calling scsi_done callback passed from midlayer when
50 invoking hostt->queuecommand() or the block la 50 invoking hostt->queuecommand() or the block layer will time it out.
51 51
52 52
53 1.2.1 Completing a scmd w/ scsi_done 53 1.2.1 Completing a scmd w/ scsi_done
54 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 54 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
55 55
56 For all non-EH commands, scsi_done() is the co 56 For all non-EH commands, scsi_done() is the completion callback. It
57 just calls blk_complete_request() to delete th 57 just calls blk_complete_request() to delete the block layer timer and
58 raise SCSI_SOFTIRQ 58 raise SCSI_SOFTIRQ
59 59
60 SCSI_SOFTIRQ handler scsi_softirq calls scsi_d 60 SCSI_SOFTIRQ handler scsi_softirq calls scsi_decide_disposition() to
61 determine what to do with the command. scsi_d 61 determine what to do with the command. scsi_decide_disposition()
62 looks at the scmd->result value and sense data 62 looks at the scmd->result value and sense data to determine what to do
63 with the command. 63 with the command.
64 64
65 - SUCCESS 65 - SUCCESS
66 66
67 scsi_finish_command() is invoked for t 67 scsi_finish_command() is invoked for the command. The
68 function does some maintenance chores 68 function does some maintenance chores and then calls
69 scsi_io_completion() to finish the I/O 69 scsi_io_completion() to finish the I/O.
70 scsi_io_completion() then notifies the 70 scsi_io_completion() then notifies the block layer on
71 the completed request by calling blk_e 71 the completed request by calling blk_end_request and
72 friends or figures out what to do with 72 friends or figures out what to do with the remainder
73 of the data in case of an error. 73 of the data in case of an error.
74 74
75 - NEEDS_RETRY 75 - NEEDS_RETRY
76 76
77 - ADD_TO_MLQUEUE 77 - ADD_TO_MLQUEUE
78 78
79 scmd is requeued to blk queue. 79 scmd is requeued to blk queue.
80 80
81 - otherwise 81 - otherwise
82 82
83 scsi_eh_scmd_add(scmd) is invoked for 83 scsi_eh_scmd_add(scmd) is invoked for the command. See
84 [1-3] for details of this function. 84 [1-3] for details of this function.
85 85
86 86
87 1.2.2 Completing a scmd w/ timeout 87 1.2.2 Completing a scmd w/ timeout
88 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 88 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
89 89
90 The timeout handler is scsi_timeout(). When a 90 The timeout handler is scsi_timeout(). When a timeout occurs, this function
91 91
92 1. invokes optional hostt->eh_timed_out() cal 92 1. invokes optional hostt->eh_timed_out() callback. Return value can
93 be one of 93 be one of
94 94
95 - SCSI_EH_RESET_TIMER !! 95 - BLK_EH_RESET_TIMER
96 This indicates that more time is requi 96 This indicates that more time is required to finish the
97 command. Timer is restarted. 97 command. Timer is restarted.
98 98
99 - SCSI_EH_NOT_HANDLED !! 99 - BLK_EH_DONE
100 eh_timed_out() callback did not handle 100 eh_timed_out() callback did not handle the command.
101 Step #2 is taken. 101 Step #2 is taken.
102 <<
103 - SCSI_EH_DONE <<
104 eh_timed_out() completed the command. <<
105 102
106 2. scsi_abort_command() is invoked to schedul 103 2. scsi_abort_command() is invoked to schedule an asynchronous abort which may
107 issue a retry scmd->allowed + 1 times. As 104 issue a retry scmd->allowed + 1 times. Asynchronous aborts are not invoked
108 for commands for which the SCSI_EH_ABORT_S 105 for commands for which the SCSI_EH_ABORT_SCHEDULED flag is set (this
109 indicates that the command already had bee 106 indicates that the command already had been aborted once, and this is a
110 retry which failed), when retries are exce 107 retry which failed), when retries are exceeded, or when the EH deadline is
111 expired. In these cases Step #3 is taken. 108 expired. In these cases Step #3 is taken.
112 109
113 3. scsi_eh_scmd_add(scmd, SCSI_EH_CANCEL_CMD) 110 3. scsi_eh_scmd_add(scmd, SCSI_EH_CANCEL_CMD) is invoked for the
114 command. See [1-4] for more information. 111 command. See [1-4] for more information.
115 112
116 1.3 Asynchronous command aborts 113 1.3 Asynchronous command aborts
117 ------------------------------- 114 -------------------------------
118 115
119 After a timeout occurs a command abort is sch 116 After a timeout occurs a command abort is scheduled from
120 scsi_abort_command(). If the abort is success 117 scsi_abort_command(). If the abort is successful the command
121 will either be retried (if the number of retr 118 will either be retried (if the number of retries is not exhausted)
122 or terminated with DID_TIME_OUT. 119 or terminated with DID_TIME_OUT.
123 120
124 Otherwise scsi_eh_scmd_add() is invoked for t 121 Otherwise scsi_eh_scmd_add() is invoked for the command.
125 See [1-4] for more information. 122 See [1-4] for more information.
126 123
127 1.4 How EH takes over 124 1.4 How EH takes over
128 --------------------- 125 ---------------------
129 126
130 scmds enter EH via scsi_eh_scmd_add(), which d 127 scmds enter EH via scsi_eh_scmd_add(), which does the following.
131 128
132 1. Links scmd->eh_entry to shost->eh_cmd_q 129 1. Links scmd->eh_entry to shost->eh_cmd_q
133 130
134 2. Sets SHOST_RECOVERY bit in shost->shost_st 131 2. Sets SHOST_RECOVERY bit in shost->shost_state
135 132
136 3. Increments shost->host_failed 133 3. Increments shost->host_failed
137 134
138 4. Wakes up SCSI EH thread if shost->host_bus 135 4. Wakes up SCSI EH thread if shost->host_busy == shost->host_failed
139 136
140 As can be seen above, once any scmd is added t 137 As can be seen above, once any scmd is added to shost->eh_cmd_q,
141 SHOST_RECOVERY shost_state bit is turned on. 138 SHOST_RECOVERY shost_state bit is turned on. This prevents any new
142 scmd to be issued from blk queue to the host; 139 scmd to be issued from blk queue to the host; eventually, all scmds on
143 the host either complete normally, fail and ge 140 the host either complete normally, fail and get added to eh_cmd_q, or
144 time out and get added to shost->eh_cmd_q. 141 time out and get added to shost->eh_cmd_q.
145 142
146 If all scmds either complete or fail, the numb 143 If all scmds either complete or fail, the number of in-flight scmds
147 becomes equal to the number of failed scmds - 144 becomes equal to the number of failed scmds - i.e. shost->host_busy ==
148 shost->host_failed. This wakes up SCSI EH thr 145 shost->host_failed. This wakes up SCSI EH thread. So, once woken up,
149 SCSI EH thread can expect that all in-flight c 146 SCSI EH thread can expect that all in-flight commands have failed and
150 are linked on shost->eh_cmd_q. 147 are linked on shost->eh_cmd_q.
151 148
152 Note that this does not mean lower layers are 149 Note that this does not mean lower layers are quiescent. If a LLDD
153 completed a scmd with error status, the LLDD a 150 completed a scmd with error status, the LLDD and lower layers are
154 assumed to forget about the scmd at that point 151 assumed to forget about the scmd at that point. However, if a scmd
155 has timed out, unless hostt->eh_timed_out() ma 152 has timed out, unless hostt->eh_timed_out() made lower layers forget
156 about the scmd, which currently no LLDD does, 153 about the scmd, which currently no LLDD does, the command is still
157 active as long as lower layers are concerned a 154 active as long as lower layers are concerned and completion could
158 occur at any time. Of course, all such comple 155 occur at any time. Of course, all such completions are ignored as the
159 timer has already expired. 156 timer has already expired.
160 157
161 We'll talk about how SCSI EH takes actions to 158 We'll talk about how SCSI EH takes actions to abort - make LLDD
162 forget about - timed out scmds later. 159 forget about - timed out scmds later.
163 160
164 161
165 2. How SCSI EH works 162 2. How SCSI EH works
166 ==================== 163 ====================
167 164
168 LLDD's can implement SCSI EH actions in one of 165 LLDD's can implement SCSI EH actions in one of the following two
169 ways. 166 ways.
170 167
171 - Fine-grained EH callbacks 168 - Fine-grained EH callbacks
172 LLDD can implement fine-grained EH cal 169 LLDD can implement fine-grained EH callbacks and let SCSI
173 midlayer drive error handling and call 170 midlayer drive error handling and call appropriate callbacks.
174 This will be discussed further in [2-1 171 This will be discussed further in [2-1].
175 172
176 - eh_strategy_handler() callback 173 - eh_strategy_handler() callback
177 This is one big callback which should 174 This is one big callback which should perform whole error
178 handling. As such, it should do all c 175 handling. As such, it should do all chores the SCSI midlayer
179 performs during recovery. This will b 176 performs during recovery. This will be discussed in [2-2].
180 177
181 Once recovery is complete, SCSI EH resumes nor 178 Once recovery is complete, SCSI EH resumes normal operation by
182 calling scsi_restart_operations(), which 179 calling scsi_restart_operations(), which
183 180
184 1. Checks if door locking is needed and locks 181 1. Checks if door locking is needed and locks door.
185 182
186 2. Clears SHOST_RECOVERY shost_state bit 183 2. Clears SHOST_RECOVERY shost_state bit
187 184
188 3. Wakes up waiters on shost->host_wait. Thi 185 3. Wakes up waiters on shost->host_wait. This occurs if someone
189 calls scsi_block_when_processing_errors() 186 calls scsi_block_when_processing_errors() on the host.
190 (*QUESTION* why is it needed? All operati 187 (*QUESTION* why is it needed? All operations will be blocked
191 anyway after it reaches blk queue.) 188 anyway after it reaches blk queue.)
192 189
193 4. Kicks queues in all devices on the host in 190 4. Kicks queues in all devices on the host in the asses
194 191
195 192
196 2.1 EH through fine-grained callbacks 193 2.1 EH through fine-grained callbacks
197 ------------------------------------- 194 -------------------------------------
198 195
199 2.1.1 Overview 196 2.1.1 Overview
200 ^^^^^^^^^^^^^^ 197 ^^^^^^^^^^^^^^
201 198
202 If eh_strategy_handler() is not present, SCSI 199 If eh_strategy_handler() is not present, SCSI midlayer takes charge
203 of driving error handling. EH's goals are two 200 of driving error handling. EH's goals are two - make LLDD, host and
204 device forget about timed out scmds and make t 201 device forget about timed out scmds and make them ready for new
205 commands. A scmd is said to be recovered if t 202 commands. A scmd is said to be recovered if the scmd is forgotten by
206 lower layers and lower layers are ready to pro 203 lower layers and lower layers are ready to process or fail the scmd
207 again. 204 again.
208 205
209 To achieve these goals, EH performs recovery a 206 To achieve these goals, EH performs recovery actions with increasing
210 severity. Some actions are performed by issui 207 severity. Some actions are performed by issuing SCSI commands and
211 others are performed by invoking one of the fo 208 others are performed by invoking one of the following fine-grained
212 hostt EH callbacks. Callbacks may be omitted 209 hostt EH callbacks. Callbacks may be omitted and omitted ones are
213 considered to fail always. 210 considered to fail always.
214 211
215 :: 212 ::
216 213
217 int (* eh_abort_handler)(struct scsi_cmnd 214 int (* eh_abort_handler)(struct scsi_cmnd *);
218 int (* eh_device_reset_handler)(struct scs 215 int (* eh_device_reset_handler)(struct scsi_cmnd *);
219 int (* eh_bus_reset_handler)(struct scsi_c 216 int (* eh_bus_reset_handler)(struct scsi_cmnd *);
220 int (* eh_host_reset_handler)(struct scsi_ 217 int (* eh_host_reset_handler)(struct scsi_cmnd *);
221 218
222 Higher-severity actions are taken only when lo 219 Higher-severity actions are taken only when lower-severity actions
223 cannot recover some of failed scmds. Also, no 220 cannot recover some of failed scmds. Also, note that failure of the
224 highest-severity action means EH failure and r 221 highest-severity action means EH failure and results in offlining of
225 all unrecovered devices. 222 all unrecovered devices.
226 223
227 During recovery, the following rules are follo 224 During recovery, the following rules are followed
228 225
229 - Recovery actions are performed on failed sc 226 - Recovery actions are performed on failed scmds on the to do list,
230 eh_work_q. If a recovery action succeeds f 227 eh_work_q. If a recovery action succeeds for a scmd, recovered
231 scmds are removed from eh_work_q. 228 scmds are removed from eh_work_q.
232 229
233 Note that single recovery action on a scmd 230 Note that single recovery action on a scmd can recover multiple
234 scmds. e.g. resetting a device recovers al 231 scmds. e.g. resetting a device recovers all failed scmds on the
235 device. 232 device.
236 233
237 - Higher severity actions are taken iff eh_wo 234 - Higher severity actions are taken iff eh_work_q is not empty after
238 lower severity actions are complete. 235 lower severity actions are complete.
239 236
240 - EH reuses failed scmds to issue commands fo 237 - EH reuses failed scmds to issue commands for recovery. For
241 timed-out scmds, SCSI EH ensures that LLDD 238 timed-out scmds, SCSI EH ensures that LLDD forgets about a scmd
242 before reusing it for EH commands. 239 before reusing it for EH commands.
243 240
244 When a scmd is recovered, the scmd is moved fr 241 When a scmd is recovered, the scmd is moved from eh_work_q to EH
245 local eh_done_q using scsi_eh_finish_cmd(). A 242 local eh_done_q using scsi_eh_finish_cmd(). After all scmds are
246 recovered (eh_work_q is empty), scsi_eh_flush_ 243 recovered (eh_work_q is empty), scsi_eh_flush_done_q() is invoked to
247 either retry or error-finish (notify upper lay 244 either retry or error-finish (notify upper layer of failure) recovered
248 scmds. 245 scmds.
249 246
250 scmds are retried iff its sdev is still online 247 scmds are retried iff its sdev is still online (not offlined during
251 EH), REQ_FAILFAST is not set and ++scmd->retri 248 EH), REQ_FAILFAST is not set and ++scmd->retries is less than
252 scmd->allowed. 249 scmd->allowed.
253 250
254 251
255 2.1.2 Flow of scmds through EH 252 2.1.2 Flow of scmds through EH
256 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 253 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
257 254
258 1. Error completion / time out 255 1. Error completion / time out
259 256
260 :ACTION: scsi_eh_scmd_add() is invoked for 257 :ACTION: scsi_eh_scmd_add() is invoked for scmd
261 258
262 - add scmd to shost->eh_cmd_q 259 - add scmd to shost->eh_cmd_q
263 - set SHOST_RECOVERY 260 - set SHOST_RECOVERY
264 - shost->host_failed++ 261 - shost->host_failed++
265 262
266 :LOCKING: shost->host_lock 263 :LOCKING: shost->host_lock
267 264
268 2. EH starts 265 2. EH starts
269 266
270 :ACTION: move all scmds to EH's local eh_w 267 :ACTION: move all scmds to EH's local eh_work_q. shost->eh_cmd_q
271 is cleared. 268 is cleared.
272 269
273 :LOCKING: shost->host_lock (not strictly n 270 :LOCKING: shost->host_lock (not strictly necessary, just for
274 consistency) 271 consistency)
275 272
276 3. scmd recovered 273 3. scmd recovered
277 274
278 :ACTION: scsi_eh_finish_cmd() is invoked t 275 :ACTION: scsi_eh_finish_cmd() is invoked to EH-finish scmd
279 276
280 - scsi_setup_cmd_retry() 277 - scsi_setup_cmd_retry()
281 - move from local eh_work_q to local e 278 - move from local eh_work_q to local eh_done_q
282 279
283 :LOCKING: none 280 :LOCKING: none
284 281
285 :CONCURRENCY: at most one thread per separ 282 :CONCURRENCY: at most one thread per separate eh_work_q to
286 keep queue manipulation lock 283 keep queue manipulation lockless
287 284
288 4. EH completes 285 4. EH completes
289 286
290 :ACTION: scsi_eh_flush_done_q() retries sc 287 :ACTION: scsi_eh_flush_done_q() retries scmds or notifies upper
291 layer of failure. May be called c 288 layer of failure. May be called concurrently but must have
292 a no more than one thread per sep 289 a no more than one thread per separate eh_work_q to
293 manipulate the queue locklessly 290 manipulate the queue locklessly
294 291
295 - scmd is removed from eh_done_q 292 - scmd is removed from eh_done_q and scmd->eh_entry is cleared
296 - if retry is necessary, scmd is 293 - if retry is necessary, scmd is requeued using
297 scsi_queue_insert() 294 scsi_queue_insert()
298 - otherwise, scsi_finish_command( 295 - otherwise, scsi_finish_command() is invoked for scmd
299 - zero shost->host_failed 296 - zero shost->host_failed
300 297
301 :LOCKING: queue or finish function perform 298 :LOCKING: queue or finish function performs appropriate locking
302 299
303 300
304 2.1.3 Flow of control 301 2.1.3 Flow of control
305 ^^^^^^^^^^^^^^^^^^^^^^ 302 ^^^^^^^^^^^^^^^^^^^^^^
306 303
307 EH through fine-grained callbacks start from 304 EH through fine-grained callbacks start from scsi_unjam_host().
308 305
309 ``scsi_unjam_host`` 306 ``scsi_unjam_host``
310 307
311 1. Lock shost->host_lock, splice_init shos 308 1. Lock shost->host_lock, splice_init shost->eh_cmd_q into local
312 eh_work_q and unlock host_lock. Note t 309 eh_work_q and unlock host_lock. Note that shost->eh_cmd_q is
313 cleared by this action. 310 cleared by this action.
314 311
315 2. Invoke scsi_eh_get_sense. 312 2. Invoke scsi_eh_get_sense.
316 313
317 ``scsi_eh_get_sense`` 314 ``scsi_eh_get_sense``
318 315
319 This action is taken for each error-co 316 This action is taken for each error-completed
320 (!SCSI_EH_CANCEL_CMD) commands without 317 (!SCSI_EH_CANCEL_CMD) commands without valid sense data. Most
321 SCSI transports/LLDDs automatically ac 318 SCSI transports/LLDDs automatically acquire sense data on
322 command failures (autosense). Autosen 319 command failures (autosense). Autosense is recommended for
323 performance reasons and as sense infor 320 performance reasons and as sense information could get out of
324 sync between occurrence of CHECK CONDI 321 sync between occurrence of CHECK CONDITION and this action.
325 322
326 Note that if autosense is not supporte 323 Note that if autosense is not supported, scmd->sense_buffer
327 contains invalid sense data when error 324 contains invalid sense data when error-completing the scmd
328 with scsi_done(). scsi_decide_disposi 325 with scsi_done(). scsi_decide_disposition() always returns
329 FAILED in such cases thus invoking SCS 326 FAILED in such cases thus invoking SCSI EH. When the scmd
330 reaches here, sense data is acquired a 327 reaches here, sense data is acquired and
331 scsi_decide_disposition() is called ag 328 scsi_decide_disposition() is called again.
332 329
333 1. Invoke scsi_request_sense() which i 330 1. Invoke scsi_request_sense() which issues REQUEST_SENSE
334 command. If fails, no action. Not 331 command. If fails, no action. Note that taking no action
335 causes higher-severity recovery to 332 causes higher-severity recovery to be taken for the scmd.
336 333
337 2. Invoke scsi_decide_disposition() on 334 2. Invoke scsi_decide_disposition() on the scmd
338 335
339 - SUCCESS 336 - SUCCESS
340 scmd->retries is set to scmd-> 337 scmd->retries is set to scmd->allowed preventing
341 scsi_eh_flush_done_q() from re 338 scsi_eh_flush_done_q() from retrying the scmd and
342 scsi_eh_finish_cmd() is invoke 339 scsi_eh_finish_cmd() is invoked.
343 340
344 - NEEDS_RETRY 341 - NEEDS_RETRY
345 scsi_eh_finish_cmd() invoked 342 scsi_eh_finish_cmd() invoked
346 343
347 - otherwise 344 - otherwise
348 No action. 345 No action.
349 346
350 3. If !list_empty(&eh_work_q), invoke scsi 347 3. If !list_empty(&eh_work_q), invoke scsi_eh_abort_cmds().
351 348
352 ``scsi_eh_abort_cmds`` 349 ``scsi_eh_abort_cmds``
353 350
354 This action is taken for each timed ou 351 This action is taken for each timed out command when
355 no_async_abort is enabled in the host 352 no_async_abort is enabled in the host template.
356 hostt->eh_abort_handler() is invoked f 353 hostt->eh_abort_handler() is invoked for each scmd. The
357 handler returns SUCCESS if it has succ 354 handler returns SUCCESS if it has succeeded to make LLDD and
358 all related hardware forget about the 355 all related hardware forget about the scmd.
359 356
360 If a timedout scmd is successfully abo 357 If a timedout scmd is successfully aborted and the sdev is
361 either offline or ready, scsi_eh_finis 358 either offline or ready, scsi_eh_finish_cmd() is invoked for
362 the scmd. Otherwise, the scmd is left 359 the scmd. Otherwise, the scmd is left in eh_work_q for
363 higher-severity actions. 360 higher-severity actions.
364 361
365 Note that both offline and ready statu 362 Note that both offline and ready status mean that the sdev is
366 ready to process new scmds, where proc 363 ready to process new scmds, where processing also implies
367 immediate failing; thus, if a sdev is 364 immediate failing; thus, if a sdev is in one of the two
368 states, no further recovery action is 365 states, no further recovery action is needed.
369 366
370 Device readiness is tested using scsi_ 367 Device readiness is tested using scsi_eh_tur() which issues
371 TEST_UNIT_READY command. Note that th 368 TEST_UNIT_READY command. Note that the scmd must have been
372 aborted successfully before reusing it 369 aborted successfully before reusing it for TEST_UNIT_READY.
373 370
374 4. If !list_empty(&eh_work_q), invoke scsi 371 4. If !list_empty(&eh_work_q), invoke scsi_eh_ready_devs()
375 372
376 ``scsi_eh_ready_devs`` 373 ``scsi_eh_ready_devs``
377 374
378 This function takes four increasingly 375 This function takes four increasingly more severe measures to
379 make failed sdevs ready for new comman 376 make failed sdevs ready for new commands.
380 377
381 1. Invoke scsi_eh_stu() 378 1. Invoke scsi_eh_stu()
382 379
383 ``scsi_eh_stu`` 380 ``scsi_eh_stu``
384 381
385 For each sdev which has failed scm 382 For each sdev which has failed scmds with valid sense data
386 of which scsi_check_sense()'s verd 383 of which scsi_check_sense()'s verdict is FAILED,
387 START_STOP_UNIT command is issued 384 START_STOP_UNIT command is issued w/ start=1. Note that
388 as we explicitly choose error-comp 385 as we explicitly choose error-completed scmds, it is known
389 that lower layers have forgotten a 386 that lower layers have forgotten about the scmd and we can
390 reuse it for STU. 387 reuse it for STU.
391 388
392 If STU succeeds and the sdev is ei 389 If STU succeeds and the sdev is either offline or ready,
393 all failed scmds on the sdev are E 390 all failed scmds on the sdev are EH-finished with
394 scsi_eh_finish_cmd(). 391 scsi_eh_finish_cmd().
395 392
396 *NOTE* If hostt->eh_abort_handler( 393 *NOTE* If hostt->eh_abort_handler() isn't implemented or
397 failed, we may still have timed ou 394 failed, we may still have timed out scmds at this point
398 and STU doesn't make lower layers 395 and STU doesn't make lower layers forget about those
399 scmds. Yet, this function EH-fini 396 scmds. Yet, this function EH-finish all scmds on the sdev
400 if STU succeeds leaving lower laye 397 if STU succeeds leaving lower layers in an inconsistent
401 state. It seems that STU action s 398 state. It seems that STU action should be taken only when
402 a sdev has no timed out scmd. 399 a sdev has no timed out scmd.
403 400
404 2. If !list_empty(&eh_work_q), invoke 401 2. If !list_empty(&eh_work_q), invoke scsi_eh_bus_device_reset().
405 402
406 ``scsi_eh_bus_device_reset`` 403 ``scsi_eh_bus_device_reset``
407 404
408 This action is very similar to scs 405 This action is very similar to scsi_eh_stu() except that,
409 instead of issuing STU, hostt->eh_ 406 instead of issuing STU, hostt->eh_device_reset_handler()
410 is used. Also, as we're not issui 407 is used. Also, as we're not issuing SCSI commands and
411 resetting clears all scmds on the 408 resetting clears all scmds on the sdev, there is no need
412 to choose error-completed scmds. 409 to choose error-completed scmds.
413 410
414 3. If !list_empty(&eh_work_q), invoke 411 3. If !list_empty(&eh_work_q), invoke scsi_eh_bus_reset()
415 412
416 ``scsi_eh_bus_reset`` 413 ``scsi_eh_bus_reset``
417 414
418 hostt->eh_bus_reset_handler() is i 415 hostt->eh_bus_reset_handler() is invoked for each channel
419 with failed scmds. If bus reset s 416 with failed scmds. If bus reset succeeds, all failed
420 scmds on all ready or offline sdev 417 scmds on all ready or offline sdevs on the channel are
421 EH-finished. 418 EH-finished.
422 419
423 4. If !list_empty(&eh_work_q), invoke 420 4. If !list_empty(&eh_work_q), invoke scsi_eh_host_reset()
424 421
425 ``scsi_eh_host_reset`` 422 ``scsi_eh_host_reset``
426 423
427 This is the last resort. hostt->e 424 This is the last resort. hostt->eh_host_reset_handler()
428 is invoked. If host reset succeed 425 is invoked. If host reset succeeds, all failed scmds on
429 all ready or offline sdevs on the 426 all ready or offline sdevs on the host are EH-finished.
430 427
431 5. If !list_empty(&eh_work_q), invoke 428 5. If !list_empty(&eh_work_q), invoke scsi_eh_offline_sdevs()
432 429
433 ``scsi_eh_offline_sdevs`` 430 ``scsi_eh_offline_sdevs``
434 431
435 Take all sdevs which still have un 432 Take all sdevs which still have unrecovered scmds offline
436 and EH-finish the scmds. 433 and EH-finish the scmds.
437 434
438 5. Invoke scsi_eh_flush_done_q(). 435 5. Invoke scsi_eh_flush_done_q().
439 436
440 ``scsi_eh_flush_done_q`` 437 ``scsi_eh_flush_done_q``
441 438
442 At this point all scmds are recove 439 At this point all scmds are recovered (or given up) and
443 put on eh_done_q by scsi_eh_finish 440 put on eh_done_q by scsi_eh_finish_cmd(). This function
444 flushes eh_done_q by either retryi 441 flushes eh_done_q by either retrying or notifying upper
445 layer of failure of the scmds. 442 layer of failure of the scmds.
446 443
447 444
448 2.2 EH through transportt->eh_strategy_handler 445 2.2 EH through transportt->eh_strategy_handler()
449 ---------------------------------------------- 446 ------------------------------------------------
450 447
451 transportt->eh_strategy_handler() is invoked i 448 transportt->eh_strategy_handler() is invoked in the place of
452 scsi_unjam_host() and it is responsible for wh 449 scsi_unjam_host() and it is responsible for whole recovery process.
453 On completion, the handler should have made lo 450 On completion, the handler should have made lower layers forget about
454 all failed scmds and either ready for new comm 451 all failed scmds and either ready for new commands or offline. Also,
455 it should perform SCSI EH maintenance chores t 452 it should perform SCSI EH maintenance chores to maintain integrity of
456 SCSI midlayer. IOW, of the steps described in 453 SCSI midlayer. IOW, of the steps described in [2-1-2], all steps
457 except for #1 must be implemented by eh_strate 454 except for #1 must be implemented by eh_strategy_handler().
458 455
459 456
460 2.2.1 Pre transportt->eh_strategy_handler() SC 457 2.2.1 Pre transportt->eh_strategy_handler() SCSI midlayer conditions
461 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 458 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
462 459
463 The following conditions are true on entry to 460 The following conditions are true on entry to the handler.
464 461
465 - Each failed scmd's eh_flags field is set ap 462 - Each failed scmd's eh_flags field is set appropriately.
466 463
467 - Each failed scmd is linked on scmd->eh_cmd_ 464 - Each failed scmd is linked on scmd->eh_cmd_q by scmd->eh_entry.
468 465
469 - SHOST_RECOVERY is set. 466 - SHOST_RECOVERY is set.
470 467
471 - shost->host_failed == shost->host_busy 468 - shost->host_failed == shost->host_busy
472 469
473 470
474 2.2.2 Post transportt->eh_strategy_handler() S 471 2.2.2 Post transportt->eh_strategy_handler() SCSI midlayer conditions
475 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 472 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
476 473
477 The following conditions must be true on exit 474 The following conditions must be true on exit from the handler.
478 475
479 - shost->host_failed is zero. 476 - shost->host_failed is zero.
480 477
481 - Each scmd is in such a state that scsi_setu 478 - Each scmd is in such a state that scsi_setup_cmd_retry() on the
482 scmd doesn't make any difference. 479 scmd doesn't make any difference.
483 480
484 - shost->eh_cmd_q is cleared. 481 - shost->eh_cmd_q is cleared.
485 482
486 - Each scmd->eh_entry is cleared. 483 - Each scmd->eh_entry is cleared.
487 484
488 - Either scsi_queue_insert() or scsi_finish_c 485 - Either scsi_queue_insert() or scsi_finish_command() is called on
489 each scmd. Note that the handler is free t 486 each scmd. Note that the handler is free to use scmd->retries and
490 ->allowed to limit the number of retries. 487 ->allowed to limit the number of retries.
491 488
492 489
493 2.2.3 Things to consider 490 2.2.3 Things to consider
494 ^^^^^^^^^^^^^^^^^^^^^^^^ 491 ^^^^^^^^^^^^^^^^^^^^^^^^
495 492
496 - Know that timed out scmds are still active 493 - Know that timed out scmds are still active on lower layers. Make
497 lower layers forget about them before doing 494 lower layers forget about them before doing anything else with
498 those scmds. 495 those scmds.
499 496
500 - For consistency, when accessing/modifying s 497 - For consistency, when accessing/modifying shost data structure,
501 grab shost->host_lock. 498 grab shost->host_lock.
502 499
503 - On completion, each failed sdev must have f 500 - On completion, each failed sdev must have forgotten about all
504 active scmds. 501 active scmds.
505 502
506 - On completion, each failed sdev must be rea 503 - On completion, each failed sdev must be ready for new commands or
507 offline. 504 offline.
508 505
509 506
510 Tejun Heo 507 Tejun Heo
511 htejun@gmail.com 508 htejun@gmail.com
512 509
513 11th September 2005 510 11th September 2005
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.