- 08 Feb, 2008 40 commits
-
-
Jonathan Brassow authored
This patch gives the ability to respond-to/record device failures that happen during read operations. It also adds the ability to read from mirror devices that are not the primary if they are in-sync. There are essentially two read paths in mirroring; the direct path and the queued path. When a read request is mapped, if the region is 'in-sync' the direct path is taken; otherwise the queued path is taken. If the direct path is taken, we must record bio information so that if the read fails we can retry it. We then discover the status of a direct read through mirror_end_io. If the read has failed, we will mark the device from which the read was attempted as failed (so we don't try to read from it again), restore the bio and try again. If the queued path is taken, we discover the results of the read from 'read_callback'. If the device failed, we will mark the device as failed and attempt the read again if there is another device where this region is known to be 'in-sync'. Signed-off-by:
Jonathan Brassow <jbrassow@redhat.com> Signed-off-by:
Alasdair G Kergon <agk@redhat.com>
-
Jonathan Brassow authored
This patch adds the ability to requeue write I/O to core device-mapper when there is a log device failure. If a write to the log produces and error, the pending writes are put on the "failures" list. Since the log is marked as failed, they will stay on the failures list until a suspend happens. Suspends come in two phases, presuspend and postsuspend. We must make sure that all the writes on the failures list are requeued in the presuspend phase (a requirement of dm core). This means that recovery must be complete (because writes may be delayed behind it) and the failures list must be requeued before we return from presuspend. The mechanisms to ensure recovery is complete (or stopped) was already in place, but needed to be moved from postsuspend to presuspend. We rely on 'flush_workqueue' to ensure that the mirror thread is complete and therefore, has requeued all writes in the failures list. Because we are using flush_workqueue, we must ensure that no additional 'queue_work' calls will produce additional I/O that we need to requeue (because once we return from presuspend, we are unable to do anything about it). 'queue_work' is called in response to the following functions: - complete_resync_work = NA, recovery is stopped - rh_dec (mirror_end_io) = NA, only calls 'queue_work' if it is ready to recover the region (recovery is stopped) or it needs to clear the region in the log* **this doesn't get called while suspending** - rh_recovery_end = NA, recovery is stopped - rh_recovery_start = NA, recovery is stopped - write_callback = 1) Writes w/o failures simply call bio_endio -> mirror_end_io -> rh_dec (see rh_dec above) 2) Writes with failures are put on the failures list and queue_work is called** ** write_callbacks don't happen during suspend ** - do_failures = NA, 'queue_work' not called if suspending - add_mirror (initialization) = NA, only done on mirror creation - queue_bio = NA, 1) delayed I/O scheduled before flush_workqueue is called. 2) No more I/Os are being issued. 3) Re-attempted READs can still be handled. (Write completions are handled through rh_dec/ write_callback - mention above - and do not use queue_bio.) Signed-off-by: -
Jonathan Brassow authored
This patch adds the calls to 'fail_mirror' if an error occurs during mirror recovery (aka resynchronization). 'fail_mirror' is responsible for recording the type of error by mirror device and ensuring an event gets raised for the purpose of notifying userspace. Signed-off-by:
-
Jonathan Brassow authored
This patch gives mirror the ability to handle device failures during normal write operations. The 'write_callback' function is called when a write completes. If all the writes failed or succeeded, we report failure or success respectively. If some of the writes failed, we call fail_mirror; which increments the error count for the device, notes the type of error encountered (DM_RAID1_WRITE_ERROR), and selects a new primary (if necessary). Note that the primary device can never change while the mirror is not in-sync (IOW, while recovery is happening.) This means that the scenario where a failed write changes the primary and gives recovery_complete a chance to misread the primary never happens. The fact that the primary can change has necessitated the change to the default_mirror field. We need to protect against reading garbage while the primary changes. We then add the bio to a new list in the mirror set, 'failures'. For every bio in the 'failures' list, we call a new function, '__bio_mark_nosync', where we mark the region 'not-in-sync' in the log and properly set the region state as, RH_NOSYNC. Userspace must also be notified of the failure. This is done by 'raising an event' (dm_table_event()). If fail_mirror is called in process context the event can be raised right away. If in interrupt context, the event is deferred to the kmirrord thread - which raises the event if 'event_waiting' is set. Backwards compatibility is maintained by ignoring errors if the DM_FEATURES_HANDLE_ERRORS flag is not present. Signed-off-by:
-
Milan Broz authored
Provided sector_t is 64 bits, reduce the in-memory footprint of the snapshot exception table by the simple method of using unused bits of the chunk number to combine consecutive entries. Signed-off-by:
Milan Broz <mbroz@redhat.com> Signed-off-by:
-
Brian Wood authored
This patch adds additional information to the status line. It is added at the end of the returned text so it will not interfere with existing implementations using this data. The addition of this information will allow for a common return interface to match that returned with the dm-raid1.c status line (with Jonathan Brassow's patches). Here is a sample of what is returned with a mirror "status" call: isw_eeaaabgfg_mirror: 0 488390920 mirror 2 8:16 8:32 3727/3727 1 AA 1 core Here's what's returned with this patch for a stripe "status" call: isw_dheeijjdej_stripe: 0 976783872 striped 2 8:16 8:32 1 AA Signed-off-by:
Brian Wood <brian.j.wood@intel.com> Signed-off-by:
-
Brian Wood authored
This patch adds the stripe_end_io function to process errors that might occur after an IO operation. As part of this there are a number of enhancements made to record and trigger events: - New atomic variable in struct stripe to record the number of errors each stripe volume device has experienced (could be used later with uevents to report back directly to userspace) - New workqueue/work struct setup to process the trigger_event function - New end_io function. It is here that testing for BIO error conditions take place. It determines the exact stripe that cause the error, records this in the new atomic variable, and calls the queue_work() function - New trigger_event function to process failure events. This calls dm_table_event() Signed-off-by:
-
Jonathan Brassow authored
If the log type is not recognised, attempt to load the module 'dm-log-<type>.ko'. Signed-off-by:
-
Milan Broz authored
Add a single-thread workqueue for each mapped device and move flushing of the lists of pushback and deferred bios to this new workqueue. Signed-off-by:
-
Milan Broz authored
dm-crypt: Use crypto ablkcipher interface Move encrypt/decrypt core to async crypto call. Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by:
-
Milan Broz authored
dm-crypt: Use crypto ablkcipher interface Prepare callback function for async crypto operation. Signed-off-by:
-
Milan Broz authored
dm-crypt: Use crypto ablkcipher interface Prepare completion for async crypto request. Signed-off-by:
-
Milan Broz authored
dm-crypt: Use crypto ablkcipher interface Introduce mempool for async crypto requests. cc->req is used mainly during synchronous operations (to prevent allocation and deallocation of the same object). Signed-off-by:
-
Milan Broz authored
dm-crypt: Use crypto ablkcipher interface Move scatterlists to separate dm_crypt_struct and pick out block processing from crypt_convert. Signed-off-by:
-
Milan Broz authored
Make io reference counting more obvious. Signed-off-by:
-
Milan Broz authored
Introduce crypt_write_io_loop(). Signed-off-by:
-
Milan Broz authored
Process write request in separate function and queue final bio through io workqueue. Signed-off-by:
-
Milan Broz authored
Add sector into dm_crypt_io instead of using local variable. Signed-off-by:
-
Alasdair G Kergon authored
Reorder kcryptd functions for clarity. Signed-off-by: -
Milan Broz authored
Rename functions to follow calling convention. Prepare write io error processing function skeleton. Signed-off-by:
-
Milan Broz authored
Simplify crypt_endio function. Signed-off-by:
-
Milan Broz authored
Move error code setting outside of crypt_dec_pending function. Use -EIO if crypt_convert_scatterlist() fails. Signed-off-by:
-
Milan Broz authored
Remove write attribute from convert_context and use bio flag instead. Signed-off-by:
-
Milan Broz authored
Move convert_context inside dm_crypt_io. Signed-off-by:
-
Alasdair G Kergon authored
A static declaration missing. Signed-off-by: -
Alasdair G Kergon authored
Drop the EXPERIMENTAL tag from well-established device-mapper targets, so the newer ones stand out better. Signed-off-by: -
Milan Broz authored
Move completion wait to separate function Signed-off-by:
-
Milan Broz authored
Change io_locking to allow processing flush in separate thread. Because we have DMF_BLOCK_IO already set, any possible new ios are queued in dm_requests now. In the case of interrupting previous wait there can be more ios queued (we unlocked io_lock for a while) but this is safe. Signed-off-by:
-
Milan Broz authored
Tidy dm_suspend function - change return value logic in dm_suspend - use atomic_read only once. - move DMF_BLOCK_IO clearing into one place Signed-off-by:
-
Milan Broz authored
Refactor deferred bio_list processing. - use separate _merge_pushback_list function - move deferred bio list pick up to flush function - use bio_list_pop instead of bio_list_get - simplify noflush flag use No real functional change in this patch. Signed-off-by:
-
Milan Broz authored
Tidy labels in alloc_dev to make later patches more clear. No functional change in this patch. Signed-off-by:
-
Andrew Morton authored
drivers/md/dm-ioctl.c:1405: warning: 'param' may be used uninitialized in this function Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Andrew Morton authored
drivers/md/dm-table.c: In function 'dm_get_device': drivers/md/dm-table.c:478: warning: 'dev' may be used uninitialized in this function Signed-off-by:
-
Andrew Morton authored
drivers/md/dm-exception-store.c: In function 'persistent_read_metadata': drivers/md/dm-exception-store.c:452: warning: 'new_snapshot' may be used uninitialized in this function Signed-off-by:
-
Daniel Walker authored
Replace semaphore with mutex. Signed-off-by:
Daniel Walker <dwalker@mvista.com> Signed-off-by:
-
Robert P. J. Day authored
Since the source file already includes the log2.h header file, it seems pointless to re-invent the necessary routine. Signed-off-by:
Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by:
-
Jun'ichi Nomura authored
"total = 0" does nothing. Signed-off-by:
Jun'ichi Nomura <j-nomura@ce.jp.nec.com> Signed-off-by:
-
Vasily Averin authored
Save some bytes. Signed-off-by:
Vasily Averin <vvs@sw.ru> Signed-off-by:
-
Paul Jimenez authored
This patch is some minor janitorish cleanup, using some macros from linux/list.h (already #included via dm.h) to improve readability. Signed-off-by:
Paul Jimenez <pj@place.org> Signed-off-by:
-
Milan Broz authored
Move compat_ioctl handling into dm-ioctl.c. Signed-off-by:
-
