linux/Documentation/scsi/scsi_eh.txt
<<
>>
Prefs
   1
   2SCSI EH
   3======================================
   4
   5 This document describes SCSI midlayer error handling infrastructure.
   6Please refer to Documentation/scsi/scsi_mid_low_api.txt for more
   7information regarding SCSI midlayer.
   8
   9TABLE OF CONTENTS
  10
  11[1] How SCSI commands travel through the midlayer and to EH
  12    [1-1] struct scsi_cmnd
  13    [1-2] How do scmd's get completed?
  14        [1-2-1] Completing a scmd w/ scsi_done
  15        [1-2-2] Completing a scmd w/ timeout
  16    [1-3] How EH takes over
  17[2] How SCSI EH works
  18    [2-1] EH through fine-grained callbacks
  19        [2-1-1] Overview
  20        [2-1-2] Flow of scmds through EH
  21        [2-1-3] Flow of control
  22    [2-2] EH through transportt->eh_strategy_handler()
  23        [2-2-1] Pre transportt->eh_strategy_handler() SCSI midlayer conditions
  24        [2-2-2] Post transportt->eh_strategy_handler() SCSI midlayer conditions
  25        [2-2-3] Things to consider
  26
  27
  28[1] How SCSI commands travel through the midlayer and to EH
  29
  30[1-1] struct scsi_cmnd
  31
  32 Each SCSI command is represented with struct scsi_cmnd (== scmd).  A
  33scmd has two list_head's to link itself into lists.  The two are
  34scmd->list and scmd->eh_entry.  The former is used for free list or
  35per-device allocated scmd list and not of much interest to this EH
  36discussion.  The latter is used for completion and EH lists and unless
  37otherwise stated scmds are always linked using scmd->eh_entry in this
  38discussion.
  39
  40
  41[1-2] How do scmd's get completed?
  42
  43 Once LLDD gets hold of a scmd, either the LLDD will complete the
  44command by calling scsi_done callback passed from midlayer when
  45invoking hostt->queuecommand() or SCSI midlayer will time it out.
  46
  47
  48[1-2-1] Completing a scmd w/ scsi_done
  49
  50 For all non-EH commands, scsi_done() is the completion callback.  It
  51does the following.
  52
  53 1. Delete timeout timer.  If it fails, it means that timeout timer
  54    has expired and is going to finish the command.  Just return.
  55
  56 2. Link scmd to per-cpu scsi_done_q using scmd->en_entry
  57
  58 3. Raise SCSI_SOFTIRQ
  59
  60 SCSI_SOFTIRQ handler scsi_softirq calls scsi_decide_disposition() to
  61determine what to do with the command.  scsi_decide_disposition()
  62looks at the scmd->result value and sense data to determine what to do
  63with the command.
  64
  65 - SUCCESS
  66        scsi_finish_command() is invoked for the command.  The
  67        function does some maintenance choirs and notify completion by
  68        calling scmd->done() callback, which, for fs requests, would
  69        be HLD completion callback - sd:sd_rw_intr, sr:rw_intr,
  70        st:st_intr.
  71
  72 - NEEDS_RETRY
  73 - ADD_TO_MLQUEUE
  74        scmd is requeued to blk queue.
  75
  76 - otherwise
  77        scsi_eh_scmd_add(scmd, 0) is invoked for the command.  See
  78        [1-3] for details of this function.
  79
  80
  81[1-2-2] Completing a scmd w/ timeout
  82
  83 The timeout handler is scsi_times_out().  When a timeout occurs, this
  84function
  85
  86 1. invokes optional hostt->eh_timed_out() callback.  Return value can
  87    be one of
  88
  89    - EH_HANDLED
  90        This indicates that eh_timed_out() dealt with the timeout.  The
  91        scmd is passed to __scsi_done() and thus linked into per-cpu
  92        scsi_done_q.  Normal command completion described in [1-2-1]
  93        follows.
  94
  95    - EH_RESET_TIMER
  96        This indicates that more time is required to finish the
  97        command.  Timer is restarted.  This action is counted as a
  98        retry and only allowed scmd->allowed + 1(!) times.  Once the
  99        limit is reached, action for EH_NOT_HANDLED is taken instead.
 100
 101        *NOTE* This action is racy as the LLDD could finish the scmd
 102        after the timeout has expired but before it's added back.  In
 103        such cases, scsi_done() would think that timeout has occurred
 104        and return without doing anything.  We lose completion and the
 105        command will time out again.
 106
 107    - EH_NOT_HANDLED
 108        This is the same as when eh_timed_out() callback doesn't exist.
 109        Step #2 is taken.
 110
 111 2. scsi_eh_scmd_add(scmd, SCSI_EH_CANCEL_CMD) is invoked for the
 112    command.  See [1-3] for more information.
 113
 114
 115[1-3] How EH takes over
 116
 117 scmds enter EH via scsi_eh_scmd_add(), which does the following.
 118
 119 1. Turns on scmd->eh_eflags as requested.  It's 0 for error
 120    completions and SCSI_EH_CANCEL_CMD for timeouts.
 121
 122 2. Links scmd->eh_entry to shost->eh_cmd_q
 123
 124 3. Sets SHOST_RECOVERY bit in shost->shost_state
 125
 126 4. Increments shost->host_failed
 127
 128 5. Wakes up SCSI EH thread if shost->host_busy == shost->host_failed
 129
 130 As can be seen above, once any scmd is added to shost->eh_cmd_q,
 131SHOST_RECOVERY shost_state bit is turned on.  This prevents any new
 132scmd to be issued from blk queue to the host; eventually, all scmds on
 133the host either complete normally, fail and get added to eh_cmd_q, or
 134time out and get added to shost->eh_cmd_q.
 135
 136 If all scmds either complete or fail, the number of in-flight scmds
 137becomes equal to the number of failed scmds - i.e. shost->host_busy ==
 138shost->host_failed.  This wakes up SCSI EH thread.  So, once woken up,
 139SCSI EH thread can expect that all in-flight commands have failed and
 140are linked on shost->eh_cmd_q.
 141
 142 Note that this does not mean lower layers are quiescent.  If a LLDD
 143completed a scmd with error status, the LLDD and lower layers are
 144assumed to forget about the scmd at that point.  However, if a scmd
 145has timed out, unless hostt->eh_timed_out() made lower layers forget
 146about the scmd, which currently no LLDD does, the command is still
 147active as long as lower layers are concerned and completion could
 148occur at any time.  Of course, all such completions are ignored as the
 149timer has already expired.
 150
 151 We'll talk about how SCSI EH takes actions to abort - make LLDD
 152forget about - timed out scmds later.
 153
 154
 155[2] How SCSI EH works
 156
 157 LLDD's can implement SCSI EH actions in one of the following two
 158ways.
 159
 160 - Fine-grained EH callbacks
 161        LLDD can implement fine-grained EH callbacks and let SCSI
 162        midlayer drive error handling and call appropriate callbacks.
 163        This will be discussed further in [2-1].
 164
 165 - eh_strategy_handler() callback
 166        This is one big callback which should perform whole error
 167        handling.  As such, it should do all choirs SCSI midlayer
 168        performs during recovery.  This will be discussed in [2-2].
 169
 170 Once recovery is complete, SCSI EH resumes normal operation by
 171calling scsi_restart_operations(), which
 172
 173 1. Checks if door locking is needed and locks door.
 174
 175 2. Clears SHOST_RECOVERY shost_state bit
 176
 177 3. Wakes up waiters on shost->host_wait.  This occurs if someone
 178    calls scsi_block_when_processing_errors() on the host.
 179    (*QUESTION* why is it needed?  All operations will be blocked
 180    anyway after it reaches blk queue.)
 181
 182 4. Kicks queues in all devices on the host in the asses
 183
 184
 185[2-1] EH through fine-grained callbacks
 186
 187[2-1-1] Overview
 188
 189 If eh_strategy_handler() is not present, SCSI midlayer takes charge
 190of driving error handling.  EH's goals are two - make LLDD, host and
 191device forget about timed out scmds and make them ready for new
 192commands.  A scmd is said to be recovered if the scmd is forgotten by
 193lower layers and lower layers are ready to process or fail the scmd
 194again.
 195
 196 To achieve these goals, EH performs recovery actions with increasing
 197severity.  Some actions are performed by issuing SCSI commands and
 198others are performed by invoking one of the following fine-grained
 199hostt EH callbacks.  Callbacks may be omitted and omitted ones are
 200considered to fail always.
 201
 202int (* eh_abort_handler)(struct scsi_cmnd *);
 203int (* eh_device_reset_handler)(struct scsi_cmnd *);
 204int (* eh_bus_reset_handler)(struct scsi_cmnd *);
 205int (* eh_host_reset_handler)(struct scsi_cmnd *);
 206
 207 Higher-severity actions are taken only when lower-severity actions
 208cannot recover some of failed scmds.  Also, note that failure of the
 209highest-severity action means EH failure and results in offlining of
 210all unrecovered devices.
 211
 212 During recovery, the following rules are followed
 213
 214 - Recovery actions are performed on failed scmds on the to do list,
 215   eh_work_q.  If a recovery action succeeds for a scmd, recovered
 216   scmds are removed from eh_work_q.
 217
 218   Note that single recovery action on a scmd can recover multiple
 219   scmds.  e.g. resetting a device recovers all failed scmds on the
 220   device.
 221
 222 - Higher severity actions are taken iff eh_work_q is not empty after
 223   lower severity actions are complete.
 224
 225 - EH reuses failed scmds to issue commands for recovery.  For
 226   timed-out scmds, SCSI EH ensures that LLDD forgets about a scmd
 227   before reusing it for EH commands.
 228
 229 When a scmd is recovered, the scmd is moved from eh_work_q to EH
 230local eh_done_q using scsi_eh_finish_cmd().  After all scmds are
 231recovered (eh_work_q is empty), scsi_eh_flush_done_q() is invoked to
 232either retry or error-finish (notify upper layer of failure) recovered
 233scmds.
 234
 235 scmds are retried iff its sdev is still online (not offlined during
 236EH), REQ_FAILFAST is not set and ++scmd->retries is less than
 237scmd->allowed.
 238
 239
 240[2-1-2] Flow of scmds through EH
 241
 242 1. Error completion / time out
 243    ACTION: scsi_eh_scmd_add() is invoked for scmd
 244        - set scmd->eh_eflags
 245        - add scmd to shost->eh_cmd_q
 246        - set SHOST_RECOVERY
 247        - shost->host_failed++
 248    LOCKING: shost->host_lock
 249
 250 2. EH starts
 251    ACTION: move all scmds to EH's local eh_work_q.  shost->eh_cmd_q
 252            is cleared.
 253    LOCKING: shost->host_lock (not strictly necessary, just for
 254             consistency)
 255
 256 3. scmd recovered
 257    ACTION: scsi_eh_finish_cmd() is invoked to EH-finish scmd
 258        - shost->host_failed--
 259        - clear scmd->eh_eflags
 260        - scsi_setup_cmd_retry()
 261        - move from local eh_work_q to local eh_done_q
 262    LOCKING: none
 263
 264 4. EH completes
 265    ACTION: scsi_eh_flush_done_q() retries scmds or notifies upper
 266            layer of failure.
 267        - scmd is removed from eh_done_q and scmd->eh_entry is cleared
 268        - if retry is necessary, scmd is requeued using
 269          scsi_queue_insert()
 270        - otherwise, scsi_finish_command() is invoked for scmd
 271    LOCKING: queue or finish function performs appropriate locking
 272
 273
 274[2-1-3] Flow of control
 275
 276 EH through fine-grained callbacks start from scsi_unjam_host().
 277
 278<<scsi_unjam_host>>
 279
 280    1. Lock shost->host_lock, splice_init shost->eh_cmd_q into local
 281       eh_work_q and unlock host_lock.  Note that shost->eh_cmd_q is
 282       cleared by this action.
 283
 284    2. Invoke scsi_eh_get_sense.
 285
 286    <<scsi_eh_get_sense>>
 287
 288        This action is taken for each error-completed
 289        (!SCSI_EH_CANCEL_CMD) commands without valid sense data.  Most
 290        SCSI transports/LLDDs automatically acquire sense data on
 291        command failures (autosense).  Autosense is recommended for
 292        performance reasons and as sense information could get out of
 293        sync between occurrence of CHECK CONDITION and this action.
 294
 295        Note that if autosense is not supported, scmd->sense_buffer
 296        contains invalid sense data when error-completing the scmd
 297        with scsi_done().  scsi_decide_disposition() always returns
 298        FAILED in such cases thus invoking SCSI EH.  When the scmd
 299        reaches here, sense data is acquired and
 300        scsi_decide_disposition() is called again.
 301
 302        1. Invoke scsi_request_sense() which issues REQUEST_SENSE
 303           command.  If fails, no action.  Note that taking no action
 304           causes higher-severity recovery to be taken for the scmd.
 305
 306        2. Invoke scsi_decide_disposition() on the scmd
 307
 308           - SUCCESS
 309                scmd->retries is set to scmd->allowed preventing
 310                scsi_eh_flush_done_q() from retrying the scmd and
 311                scsi_eh_finish_cmd() is invoked.
 312
 313           - NEEDS_RETRY
 314                scsi_eh_finish_cmd() invoked
 315
 316           - otherwise
 317                No action.
 318
 319    3. If !list_empty(&eh_work_q), invoke scsi_eh_abort_cmds().
 320
 321    <<scsi_eh_abort_cmds>>
 322
 323        This action is taken for each timed out command.
 324        hostt->eh_abort_handler() is invoked for each scmd.  The
 325        handler returns SUCCESS if it has succeeded to make LLDD and
 326        all related hardware forget about the scmd.
 327
 328        If a timedout scmd is successfully aborted and the sdev is
 329        either offline or ready, scsi_eh_finish_cmd() is invoked for
 330        the scmd.  Otherwise, the scmd is left in eh_work_q for
 331        higher-severity actions.
 332
 333        Note that both offline and ready status mean that the sdev is
 334        ready to process new scmds, where processing also implies
 335        immediate failing; thus, if a sdev is in one of the two
 336        states, no further recovery action is needed.
 337
 338        Device readiness is tested using scsi_eh_tur() which issues
 339        TEST_UNIT_READY command.  Note that the scmd must have been
 340        aborted successfully before reusing it for TEST_UNIT_READY.
 341
 342    4. If !list_empty(&eh_work_q), invoke scsi_eh_ready_devs()
 343
 344    <<scsi_eh_ready_devs>>
 345
 346        This function takes four increasingly more severe measures to
 347        make failed sdevs ready for new commands.
 348
 349        1. Invoke scsi_eh_stu()
 350
 351        <<scsi_eh_stu>>
 352
 353            For each sdev which has failed scmds with valid sense data
 354            of which scsi_check_sense()'s verdict is FAILED,
 355            START_STOP_UNIT command is issued w/ start=1.  Note that
 356            as we explicitly choose error-completed scmds, it is known
 357            that lower layers have forgotten about the scmd and we can
 358            reuse it for STU.
 359
 360            If STU succeeds and the sdev is either offline or ready,
 361            all failed scmds on the sdev are EH-finished with
 362            scsi_eh_finish_cmd().
 363
 364            *NOTE* If hostt->eh_abort_handler() isn't implemented or
 365            failed, we may still have timed out scmds at this point
 366            and STU doesn't make lower layers forget about those
 367            scmds.  Yet, this function EH-finish all scmds on the sdev
 368            if STU succeeds leaving lower layers in an inconsistent
 369            state.  It seems that STU action should be taken only when
 370            a sdev has no timed out scmd.
 371
 372        2. If !list_empty(&eh_work_q), invoke scsi_eh_bus_device_reset().
 373
 374        <<scsi_eh_bus_device_reset>>
 375
 376            This action is very similar to scsi_eh_stu() except that,
 377            instead of issuing STU, hostt->eh_device_reset_handler()
 378            is used.  Also, as we're not issuing SCSI commands and
 379            resetting clears all scmds on the sdev, there is no need
 380            to choose error-completed scmds.
 381
 382        3. If !list_empty(&eh_work_q), invoke scsi_eh_bus_reset()
 383
 384        <<scsi_eh_bus_reset>>
 385
 386            hostt->eh_bus_reset_handler() is invoked for each channel
 387            with failed scmds.  If bus reset succeeds, all failed
 388            scmds on all ready or offline sdevs on the channel are
 389            EH-finished.
 390
 391        4. If !list_empty(&eh_work_q), invoke scsi_eh_host_reset()
 392
 393        <<scsi_eh_host_reset>>
 394
 395            This is the last resort.  hostt->eh_host_reset_handler()
 396            is invoked.  If host reset succeeds, all failed scmds on
 397            all ready or offline sdevs on the host are EH-finished.
 398
 399        5. If !list_empty(&eh_work_q), invoke scsi_eh_offline_sdevs()
 400
 401        <<scsi_eh_offline_sdevs>>
 402
 403            Take all sdevs which still have unrecovered scmds offline
 404            and EH-finish the scmds.
 405
 406    5. Invoke scsi_eh_flush_done_q().
 407
 408        <<scsi_eh_flush_done_q>>
 409
 410            At this point all scmds are recovered (or given up) and
 411            put on eh_done_q by scsi_eh_finish_cmd().  This function
 412            flushes eh_done_q by either retrying or notifying upper
 413            layer of failure of the scmds.
 414
 415
 416[2-2] EH through transportt->eh_strategy_handler()
 417
 418 transportt->eh_strategy_handler() is invoked in the place of
 419scsi_unjam_host() and it is responsible for whole recovery process.
 420On completion, the handler should have made lower layers forget about
 421all failed scmds and either ready for new commands or offline.  Also,
 422it should perform SCSI EH maintenance choirs to maintain integrity of
 423SCSI midlayer.  IOW, of the steps described in [2-1-2], all steps
 424except for #1 must be implemented by eh_strategy_handler().
 425
 426
 427[2-2-1] Pre transportt->eh_strategy_handler() SCSI midlayer conditions
 428
 429 The following conditions are true on entry to the handler.
 430
 431 - Each failed scmd's eh_flags field is set appropriately.
 432
 433 - Each failed scmd is linked on scmd->eh_cmd_q by scmd->eh_entry.
 434
 435 - SHOST_RECOVERY is set.
 436
 437 - shost->host_failed == shost->host_busy
 438
 439
 440[2-2-2] Post transportt->eh_strategy_handler() SCSI midlayer conditions
 441
 442 The following conditions must be true on exit from the handler.
 443
 444 - shost->host_failed is zero.
 445
 446 - Each scmd's eh_eflags field is cleared.
 447
 448 - Each scmd is in such a state that scsi_setup_cmd_retry() on the
 449   scmd doesn't make any difference.
 450
 451 - shost->eh_cmd_q is cleared.
 452
 453 - Each scmd->eh_entry is cleared.
 454
 455 - Either scsi_queue_insert() or scsi_finish_command() is called on
 456   each scmd.  Note that the handler is free to use scmd->retries and
 457   ->allowed to limit the number of retries.
 458
 459
 460[2-2-3] Things to consider
 461
 462 - Know that timed out scmds are still active on lower layers.  Make
 463   lower layers forget about them before doing anything else with
 464   those scmds.
 465
 466 - For consistency, when accessing/modifying shost data structure,
 467   grab shost->host_lock.
 468
 469 - On completion, each failed sdev must have forgotten about all
 470   active scmds.
 471
 472 - On completion, each failed sdev must be ready for new commands or
 473   offline.
 474
 475
 476--
 477Tejun Heo
 478htejun@gmail.com
 47911th September 2005
 480
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.