diff options
| -rw-r--r-- | drivers/block/drbd/drbd_actlog.c | 2 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_bitmap.c | 4 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_int.h | 17 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_main.c | 20 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_nl.c | 4 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_req.c | 6 | ||||
| -rw-r--r-- | drivers/block/drbd/drbd_worker.c | 4 |
7 files changed, 42 insertions, 15 deletions
diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c index e54e31b02b88..6ace11e9e5a1 100644 --- a/drivers/block/drbd/drbd_actlog.c +++ b/drivers/block/drbd/drbd_actlog.c @@ -411,7 +411,7 @@ w_al_write_transaction(struct drbd_conf *mdev, struct drbd_work *w, int unused) + mdev->ldev->md.al_offset + mdev->al_tr_pos; if (!drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) - drbd_chk_io_error(mdev, 1, true); + drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); if (++mdev->al_tr_pos > div_ceil(mdev->act_log->nr_elements, AL_EXTENTS_PT)) diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c index fcb956bb4b4c..ba91b408abad 100644 --- a/drivers/block/drbd/drbd_bitmap.c +++ b/drivers/block/drbd/drbd_bitmap.c @@ -1096,7 +1096,7 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned flags, unsigned lazy_w if (ctx->error) { dev_alert(DEV, "we had at least one MD IO ERROR during bitmap IO\n"); - drbd_chk_io_error(mdev, 1, true); + drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); err = -EIO; /* ctx->error ? */ } @@ -1212,7 +1212,7 @@ int drbd_bm_write_page(struct drbd_conf *mdev, unsigned int idx) __must_hold(loc wait_until_done_or_disk_failure(mdev, mdev->ldev, &ctx->done); if (ctx->error) - drbd_chk_io_error(mdev, 1, true); + drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); /* that should force detach, so the in memory bitmap will be * gone in a moment as well. */ diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h index 356a6e5b4415..79c69ebb0653 100644 --- a/drivers/block/drbd/drbd_int.h +++ b/drivers/block/drbd/drbd_int.h @@ -832,6 +832,7 @@ enum { BITMAP_IO_QUEUED, /* Started bitmap IO */ GO_DISKLESS, /* Disk is being detached, on io-error or admin request. */ WAS_IO_ERROR, /* Local disk failed returned IO error */ + FORCE_DETACH, /* Force-detach from local disk, aborting any pending local IO */ RESYNC_AFTER_NEG, /* Resync after online grow after the attach&negotiate finished. */ NET_CONGESTED, /* The data socket is congested */ @@ -1838,12 +1839,20 @@ static inline int drbd_request_state(struct drbd_conf *mdev, return _drbd_request_state(mdev, mask, val, CS_VERBOSE + CS_ORDERED); } +enum drbd_force_detach_flags { + DRBD_IO_ERROR, + DRBD_META_IO_ERROR, + DRBD_FORCE_DETACH, +}; + #define __drbd_chk_io_error(m,f) __drbd_chk_io_error_(m,f, __func__) -static inline void __drbd_chk_io_error_(struct drbd_conf *mdev, int forcedetach, const char *where) +static inline void __drbd_chk_io_error_(struct drbd_conf *mdev, + enum drbd_force_detach_flags forcedetach, + const char *where) { switch (mdev->ldev->dc.on_io_error) { case EP_PASS_ON: - if (!forcedetach) { + if (forcedetach == DRBD_IO_ERROR) { if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Local IO failed in %s.\n", where); if (mdev->state.disk > D_INCONSISTENT) @@ -1854,6 +1863,8 @@ static inline void __drbd_chk_io_error_(struct drbd_conf *mdev, int forcedetach, case EP_DETACH: case EP_CALL_HELPER: set_bit(WAS_IO_ERROR, &mdev->flags); + if (forcedetach == DRBD_FORCE_DETACH) + set_bit(FORCE_DETACH, &mdev->flags); if (mdev->state.disk > D_FAILED) { _drbd_set_state(_NS(mdev, disk, D_FAILED), CS_HARD, NULL); dev_err(DEV, @@ -1873,7 +1884,7 @@ static inline void __drbd_chk_io_error_(struct drbd_conf *mdev, int forcedetach, */ #define drbd_chk_io_error(m,e,f) drbd_chk_io_error_(m,e,f, __func__) static inline void drbd_chk_io_error_(struct drbd_conf *mdev, - int error, int forcedetach, const char *where) + int error, enum drbd_force_detach_flags forcedetach, const char *where) { if (error) { unsigned long flags; diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c index 920ede2829d6..5bebe8d8ace3 100644 --- a/drivers/block/drbd/drbd_main.c +++ b/drivers/block/drbd/drbd_main.c @@ -1630,9 +1630,21 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os, eh = mdev->ldev->dc.on_io_error; was_io_error = test_and_clear_bit(WAS_IO_ERROR, &mdev->flags); - /* Immediately allow completion of all application IO, that waits - for completion from the local disk. */ - tl_abort_disk_io(mdev); + /* Immediately allow completion of all application IO, + * that waits for completion from the local disk, + * if this was a force-detach due to disk_timeout + * or administrator request (drbdsetup detach --force). + * Do NOT abort otherwise. + * Aborting local requests may cause serious problems, + * if requests are completed to upper layers already, + * and then later the already submitted local bio completes. + * This can cause DMA into former bio pages that meanwhile + * have been re-used for other things. + * So aborting local requests may cause crashes, + * or even worse, silent data corruption. + */ + if (test_and_clear_bit(FORCE_DETACH, &mdev->flags)) + tl_abort_disk_io(mdev); /* current state still has to be D_FAILED, * there is only one way out: to D_DISKLESS, @@ -3870,7 +3882,7 @@ void drbd_md_sync(struct drbd_conf *mdev) if (!drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) { /* this was a try anyways ... */ dev_err(DEV, "meta data update failed!\n"); - drbd_chk_io_error(mdev, 1, true); + drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); } /* Update mdev->ldev->md.la_size_sect, diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c index 6d4de6a72e80..40a1c4f07190 100644 --- a/drivers/block/drbd/drbd_nl.c +++ b/drivers/block/drbd/drbd_nl.c @@ -950,6 +950,9 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp * to realize a "hot spare" feature (not that I'd recommend that) */ wait_event(mdev->misc_wait, !atomic_read(&mdev->local_cnt)); + /* make sure there is no leftover from previous force-detach attempts */ + clear_bit(FORCE_DETACH, &mdev->flags); + /* allocation not in the IO path, cqueue thread context */ nbc = kzalloc(sizeof(struct drbd_backing_dev), GFP_KERNEL); if (!nbc) { @@ -1345,6 +1348,7 @@ static int drbd_nl_detach(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, } if (dt.detach_force) { + set_bit(FORCE_DETACH, &mdev->flags); drbd_force_state(mdev, NS(disk, D_FAILED)); reply->ret_code = SS_SUCCESS; goto out; diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c index 8e93a6ac9bb6..1f4b2dbb7d4a 100644 --- a/drivers/block/drbd/drbd_req.c +++ b/drivers/block/drbd/drbd_req.c @@ -455,7 +455,7 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what, req->rq_state |= RQ_LOCAL_COMPLETED; req->rq_state &= ~RQ_LOCAL_PENDING; - __drbd_chk_io_error(mdev, false); + __drbd_chk_io_error(mdev, DRBD_IO_ERROR); _req_may_be_done_not_susp(req, m); break; @@ -477,7 +477,7 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what, break; } - __drbd_chk_io_error(mdev, false); + __drbd_chk_io_error(mdev, DRBD_IO_ERROR); goto_queue_for_net_read: @@ -1275,7 +1275,7 @@ void request_timer_fn(unsigned long data) time_after(now, req->start_time + dt) && !time_in_range(now, mdev->last_reattach_jif, mdev->last_reattach_jif + dt)) { dev_warn(DEV, "Local backing device failed to meet the disk-timeout\n"); - __drbd_chk_io_error(mdev, 1); + __drbd_chk_io_error(mdev, DRBD_FORCE_DETACH); } nt = (time_after(now, req->start_time + et) ? now : req->start_time) + et; spin_unlock_irq(&mdev->req_lock); diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c index 620c70ff2231..a35393f2fd1b 100644 --- a/drivers/block/drbd/drbd_worker.c +++ b/drivers/block/drbd/drbd_worker.c @@ -111,7 +111,7 @@ void drbd_endio_read_sec_final(struct drbd_epoch_entry *e) __releases(local) if (list_empty(&mdev->read_ee)) wake_up(&mdev->ee_wait); if (test_bit(__EE_WAS_ERROR, &e->flags)) - __drbd_chk_io_error(mdev, false); + __drbd_chk_io_error(mdev, DRBD_IO_ERROR); spin_unlock_irqrestore(&mdev->req_lock, flags); drbd_queue_work(&mdev->data.work, &e->w); @@ -154,7 +154,7 @@ static void drbd_endio_write_sec_final(struct drbd_epoch_entry *e) __releases(lo : list_empty(&mdev->active_ee); if (test_bit(__EE_WAS_ERROR, &e->flags)) - __drbd_chk_io_error(mdev, false); + __drbd_chk_io_error(mdev, DRBD_IO_ERROR); spin_unlock_irqrestore(&mdev->req_lock, flags); if (is_syncer_req) |