--- zzzz-none-000/linux-3.10.107/drivers/block/drbd/drbd_worker.c 2017-06-27 09:49:32.000000000 +0000 +++ scorpion-7490-727/linux-3.10.107/drivers/block/drbd/drbd_worker.c 2021-02-04 17:41:59.000000000 +0000 @@ -21,7 +21,7 @@ along with drbd; see the file COPYING. If not, write to the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. - */ +*/ #include #include @@ -36,16 +36,17 @@ #include #include "drbd_int.h" +#include "drbd_protocol.h" #include "drbd_req.h" -static int w_make_ov_request(struct drbd_work *w, int cancel); - +static int make_ov_request(struct drbd_device *, int); +static int make_resync_request(struct drbd_device *, int); /* endio handlers: - * drbd_md_io_complete (defined here) + * drbd_md_endio (defined here) * drbd_request_endio (defined here) * drbd_peer_request_endio (defined here) - * bm_async_io_complete (defined in drbd_bitmap.c) + * drbd_bm_endio (defined in drbd_bitmap.c) * * For all these callbacks, note the following: * The callbacks will be called in irq context by the IDE drivers, @@ -64,15 +65,12 @@ /* used for synchronous meta data and bitmap IO * submitted by drbd_md_sync_page_io() */ -void drbd_md_io_complete(struct bio *bio, int error) +void drbd_md_endio(struct bio *bio) { - struct drbd_md_io *md_io; - struct drbd_conf *mdev; + struct drbd_device *device; - md_io = (struct drbd_md_io *)bio->bi_private; - mdev = container_of(md_io, struct drbd_conf, md_io); - - md_io->error = error; + device = bio->bi_private; + device->md_io.error = bio->bi_error; /* We grabbed an extra reference in _drbd_md_sync_page_io() to be able * to timeout on the lower level device, and eventually detach from it. @@ -83,43 +81,45 @@ * Make sure we first drop the reference, and only then signal * completion, or we may (in drbd_al_read_log()) cycle so fast into the * next drbd_md_sync_page_io(), that we trigger the - * ASSERT(atomic_read(&mdev->md_io_in_use) == 1) there. + * ASSERT(atomic_read(&device->md_io_in_use) == 1) there. */ - drbd_md_put_buffer(mdev); - md_io->done = 1; - wake_up(&mdev->misc_wait); + drbd_md_put_buffer(device); + device->md_io.done = 1; + wake_up(&device->misc_wait); bio_put(bio); - if (mdev->ldev) /* special case: drbd_md_read() during drbd_adm_attach() */ - put_ldev(mdev); + if (device->ldev) /* special case: drbd_md_read() during drbd_adm_attach() */ + put_ldev(device); } /* reads on behalf of the partner, * "submitted" by the receiver */ -void drbd_endio_read_sec_final(struct drbd_peer_request *peer_req) __releases(local) +static void drbd_endio_read_sec_final(struct drbd_peer_request *peer_req) __releases(local) { unsigned long flags = 0; - struct drbd_conf *mdev = peer_req->w.mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; - spin_lock_irqsave(&mdev->tconn->req_lock, flags); - mdev->read_cnt += peer_req->i.size >> 9; + spin_lock_irqsave(&device->resource->req_lock, flags); + device->read_cnt += peer_req->i.size >> 9; list_del(&peer_req->w.list); - if (list_empty(&mdev->read_ee)) - wake_up(&mdev->ee_wait); + if (list_empty(&device->read_ee)) + wake_up(&device->ee_wait); if (test_bit(__EE_WAS_ERROR, &peer_req->flags)) - __drbd_chk_io_error(mdev, DRBD_READ_ERROR); - spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + __drbd_chk_io_error(device, DRBD_READ_ERROR); + spin_unlock_irqrestore(&device->resource->req_lock, flags); - drbd_queue_work(&mdev->tconn->sender_work, &peer_req->w); - put_ldev(mdev); + drbd_queue_work(&peer_device->connection->sender_work, &peer_req->w); + put_ldev(device); } /* writes on behalf of the partner, or resync writes, * "submitted" by the receiver, final stage. */ -static void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(local) +void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(local) { unsigned long flags = 0; - struct drbd_conf *mdev = peer_req->w.mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; struct drbd_interval i; int do_wake; u64 block_id; @@ -132,64 +132,58 @@ i = peer_req->i; do_al_complete_io = peer_req->flags & EE_CALL_AL_COMPLETE_IO; block_id = peer_req->block_id; + peer_req->flags &= ~EE_CALL_AL_COMPLETE_IO; - spin_lock_irqsave(&mdev->tconn->req_lock, flags); - mdev->writ_cnt += peer_req->i.size >> 9; - list_move_tail(&peer_req->w.list, &mdev->done_ee); + spin_lock_irqsave(&device->resource->req_lock, flags); + device->writ_cnt += peer_req->i.size >> 9; + list_move_tail(&peer_req->w.list, &device->done_ee); /* * Do not remove from the write_requests tree here: we did not send the * Ack yet and did not wake possibly waiting conflicting requests. * Removed from the tree from "drbd_process_done_ee" within the - * appropriate w.cb (e_end_block/e_end_resync_block) or from + * appropriate dw.cb (e_end_block/e_end_resync_block) or from * _drbd_clear_done_ee. */ - do_wake = list_empty(block_id == ID_SYNCER ? &mdev->sync_ee : &mdev->active_ee); + do_wake = list_empty(block_id == ID_SYNCER ? &device->sync_ee : &device->active_ee); - if (test_bit(__EE_WAS_ERROR, &peer_req->flags)) - __drbd_chk_io_error(mdev, DRBD_WRITE_ERROR); - spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + /* FIXME do we want to detach for failed REQ_DISCARD? + * ((peer_req->flags & (EE_WAS_ERROR|EE_IS_TRIM)) == EE_WAS_ERROR) */ + if (peer_req->flags & EE_WAS_ERROR) + __drbd_chk_io_error(device, DRBD_WRITE_ERROR); + spin_unlock_irqrestore(&device->resource->req_lock, flags); if (block_id == ID_SYNCER) - drbd_rs_complete_io(mdev, i.sector); + drbd_rs_complete_io(device, i.sector); if (do_wake) - wake_up(&mdev->ee_wait); + wake_up(&device->ee_wait); if (do_al_complete_io) - drbd_al_complete_io(mdev, &i); + drbd_al_complete_io(device, &i); - wake_asender(mdev->tconn); - put_ldev(mdev); + wake_asender(peer_device->connection); + put_ldev(device); } /* writes on behalf of the partner, or resync writes, * "submitted" by the receiver. */ -void drbd_peer_request_endio(struct bio *bio, int error) +void drbd_peer_request_endio(struct bio *bio) { struct drbd_peer_request *peer_req = bio->bi_private; - struct drbd_conf *mdev = peer_req->w.mdev; - int uptodate = bio_flagged(bio, BIO_UPTODATE); + struct drbd_device *device = peer_req->peer_device->device; int is_write = bio_data_dir(bio) == WRITE; + int is_discard = !!(bio->bi_rw & REQ_DISCARD); - if (error && __ratelimit(&drbd_ratelimit_state)) - dev_warn(DEV, "%s: error=%d s=%llus\n", - is_write ? "write" : "read", error, + if (bio->bi_error && __ratelimit(&drbd_ratelimit_state)) + drbd_warn(device, "%s: error=%d s=%llus\n", + is_write ? (is_discard ? "discard" : "write") + : "read", bio->bi_error, (unsigned long long)peer_req->i.sector); - if (!error && !uptodate) { - if (__ratelimit(&drbd_ratelimit_state)) - dev_warn(DEV, "%s: setting error to -EIO s=%llus\n", - is_write ? "write" : "read", - (unsigned long long)peer_req->i.sector); - /* strange behavior of some lower level drivers... - * fail the request by clearing the uptodate flag, - * but do not return any error?! */ - error = -EIO; - } - if (error) + if (bio->bi_error) set_bit(__EE_WAS_ERROR, &peer_req->flags); bio_put(bio); /* no need for the bio anymore */ @@ -203,24 +197,13 @@ /* read, readA or write requests on R_PRIMARY coming from drbd_make_request */ -void drbd_request_endio(struct bio *bio, int error) +void drbd_request_endio(struct bio *bio) { unsigned long flags; struct drbd_request *req = bio->bi_private; - struct drbd_conf *mdev = req->w.mdev; + struct drbd_device *device = req->device; struct bio_and_error m; enum drbd_req_event what; - int uptodate = bio_flagged(bio, BIO_UPTODATE); - - if (!error && !uptodate) { - dev_warn(DEV, "p %s: setting error to -EIO\n", - bio_data_dir(bio) == WRITE ? "write" : "read"); - /* strange behavior of some lower level drivers... - * fail the request by clearing the uptodate flag, - * but do not return any error?! */ - error = -EIO; - } - /* If this request was aborted locally before, * but now was completed "successfully", @@ -252,15 +235,20 @@ */ if (unlikely(req->rq_state & RQ_LOCAL_ABORTED)) { if (__ratelimit(&drbd_ratelimit_state)) - dev_emerg(DEV, "delayed completion of aborted local request; disk-timeout may be too aggressive\n"); + drbd_emerg(device, "delayed completion of aborted local request; disk-timeout may be too aggressive\n"); - if (!error) + if (!bio->bi_error) panic("possible random memory corruption caused by delayed completion of aborted local request\n"); } /* to avoid recursion in __req_mod */ - if (unlikely(error)) { - what = (bio_data_dir(bio) == WRITE) + if (unlikely(bio->bi_error)) { + if (bio->bi_rw & REQ_DISCARD) + what = (bio->bi_error == -EOPNOTSUPP) + ? DISCARD_COMPLETED_NOTSUPP + : DISCARD_COMPLETED_WITH_ERROR; + else + what = (bio_data_dir(bio) == WRITE) ? WRITE_COMPLETED_WITH_ERROR : (bio_rw(bio) == READ) ? READ_COMPLETED_WITH_ERROR @@ -269,20 +257,19 @@ what = COMPLETED_OK; bio_put(req->private_bio); - req->private_bio = ERR_PTR(error); + req->private_bio = ERR_PTR(bio->bi_error); /* not req_mod(), we need irqsave here! */ - spin_lock_irqsave(&mdev->tconn->req_lock, flags); + spin_lock_irqsave(&device->resource->req_lock, flags); __req_mod(req, what, &m); - spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); - put_ldev(mdev); + spin_unlock_irqrestore(&device->resource->req_lock, flags); + put_ldev(device); if (m.bio) - complete_master_bio(mdev, &m); + complete_master_bio(device, &m); } -void drbd_csum_ee(struct drbd_conf *mdev, struct crypto_hash *tfm, - struct drbd_peer_request *peer_req, void *digest) +void drbd_csum_ee(struct crypto_hash *tfm, struct drbd_peer_request *peer_req, void *digest) { struct hash_desc desc; struct scatterlist sg; @@ -309,12 +296,12 @@ crypto_hash_final(&desc, digest); } -void drbd_csum_bio(struct drbd_conf *mdev, struct crypto_hash *tfm, struct bio *bio, void *digest) +void drbd_csum_bio(struct crypto_hash *tfm, struct bio *bio, void *digest) { struct hash_desc desc; struct scatterlist sg; - struct bio_vec *bvec; - int i; + struct bio_vec bvec; + struct bvec_iter iter; desc.tfm = tfm; desc.flags = 0; @@ -322,8 +309,8 @@ sg_init_table(&sg, 1); crypto_hash_init(&desc); - bio_for_each_segment(bvec, bio, i) { - sg_set_page(&sg, bvec->bv_page, bvec->bv_len, bvec->bv_offset); + bio_for_each_segment(bvec, bio, iter) { + sg_set_page(&sg, bvec.bv_page, bvec.bv_len, bvec.bv_offset); crypto_hash_update(&desc, &sg, sg.length); } crypto_hash_final(&desc, digest); @@ -333,7 +320,8 @@ static int w_e_send_csum(struct drbd_work *w, int cancel) { struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; int digest_size; void *digest; int err = 0; @@ -344,89 +332,89 @@ if (unlikely((peer_req->flags & EE_WAS_ERROR) != 0)) goto out; - digest_size = crypto_hash_digestsize(mdev->tconn->csums_tfm); + digest_size = crypto_hash_digestsize(peer_device->connection->csums_tfm); digest = kmalloc(digest_size, GFP_NOIO); if (digest) { sector_t sector = peer_req->i.sector; unsigned int size = peer_req->i.size; - drbd_csum_ee(mdev, mdev->tconn->csums_tfm, peer_req, digest); + drbd_csum_ee(peer_device->connection->csums_tfm, peer_req, digest); /* Free peer_req and pages before send. * In case we block on congestion, we could otherwise run into * some distributed deadlock, if the other side blocks on * congestion as well, because our receiver blocks in * drbd_alloc_pages due to pp_in_use > max_buffers. */ - drbd_free_peer_req(mdev, peer_req); + drbd_free_peer_req(device, peer_req); peer_req = NULL; - inc_rs_pending(mdev); - err = drbd_send_drequest_csum(mdev, sector, size, + inc_rs_pending(device); + err = drbd_send_drequest_csum(peer_device, sector, size, digest, digest_size, P_CSUM_RS_REQUEST); kfree(digest); } else { - dev_err(DEV, "kmalloc() of digest failed.\n"); + drbd_err(device, "kmalloc() of digest failed.\n"); err = -ENOMEM; } out: if (peer_req) - drbd_free_peer_req(mdev, peer_req); + drbd_free_peer_req(device, peer_req); if (unlikely(err)) - dev_err(DEV, "drbd_send_drequest(..., csum) failed\n"); + drbd_err(device, "drbd_send_drequest(..., csum) failed\n"); return err; } #define GFP_TRY (__GFP_HIGHMEM | __GFP_NOWARN) -static int read_for_csum(struct drbd_conf *mdev, sector_t sector, int size) +static int read_for_csum(struct drbd_peer_device *peer_device, sector_t sector, int size) { + struct drbd_device *device = peer_device->device; struct drbd_peer_request *peer_req; - if (!get_ldev(mdev)) + if (!get_ldev(device)) return -EIO; - if (drbd_rs_should_slow_down(mdev, sector)) - goto defer; - /* GFP_TRY, because if there is no memory available right now, this may * be rescheduled for later. It is "only" background resync, after all. */ - peer_req = drbd_alloc_peer_req(mdev, ID_SYNCER /* unused */, sector, - size, GFP_TRY); + peer_req = drbd_alloc_peer_req(peer_device, ID_SYNCER /* unused */, sector, + size, true /* has real payload */, GFP_TRY); if (!peer_req) goto defer; peer_req->w.cb = w_e_send_csum; - spin_lock_irq(&mdev->tconn->req_lock); - list_add(&peer_req->w.list, &mdev->read_ee); - spin_unlock_irq(&mdev->tconn->req_lock); + spin_lock_irq(&device->resource->req_lock); + list_add_tail(&peer_req->w.list, &device->read_ee); + spin_unlock_irq(&device->resource->req_lock); - atomic_add(size >> 9, &mdev->rs_sect_ev); - if (drbd_submit_peer_request(mdev, peer_req, READ, DRBD_FAULT_RS_RD) == 0) + atomic_add(size >> 9, &device->rs_sect_ev); + if (drbd_submit_peer_request(device, peer_req, READ, DRBD_FAULT_RS_RD) == 0) return 0; /* If it failed because of ENOMEM, retry should help. If it failed * because bio_add_page failed (probably broken lower level driver), * retry may or may not help. * If it does not, you may need to force disconnect. */ - spin_lock_irq(&mdev->tconn->req_lock); + spin_lock_irq(&device->resource->req_lock); list_del(&peer_req->w.list); - spin_unlock_irq(&mdev->tconn->req_lock); + spin_unlock_irq(&device->resource->req_lock); - drbd_free_peer_req(mdev, peer_req); + drbd_free_peer_req(device, peer_req); defer: - put_ldev(mdev); + put_ldev(device); return -EAGAIN; } int w_resync_timer(struct drbd_work *w, int cancel) { - struct drbd_conf *mdev = w->mdev; - switch (mdev->state.conn) { + struct drbd_device *device = + container_of(w, struct drbd_device, resync_work); + + switch (device->state.conn) { case C_VERIFY_S: - w_make_ov_request(w, cancel); + make_ov_request(device, cancel); break; case C_SYNC_TARGET: - w_make_resync_request(w, cancel); + make_resync_request(device, cancel); break; } @@ -435,10 +423,11 @@ void resync_timer_fn(unsigned long data) { - struct drbd_conf *mdev = (struct drbd_conf *) data; + struct drbd_device *device = (struct drbd_device *) data; - if (list_empty(&mdev->resync_work.list)) - drbd_queue_work(&mdev->tconn->sender_work, &mdev->resync_work); + drbd_queue_work_if_unqueued( + &first_peer_device(device)->connection->sender_work, + &device->resync_work); } static void fifo_set(struct fifo_buffer *fb, int value) @@ -485,35 +474,31 @@ return fb; } -static int drbd_rs_controller(struct drbd_conf *mdev) +static int drbd_rs_controller(struct drbd_device *device, unsigned int sect_in) { struct disk_conf *dc; - unsigned int sect_in; /* Number of sectors that came in since the last turn */ - unsigned int want; /* The number of sectors we want in the proxy */ + unsigned int want; /* The number of sectors we want in-flight */ int req_sect; /* Number of sectors to request in this turn */ - int correction; /* Number of sectors more we need in the proxy*/ + int correction; /* Number of sectors more we need in-flight */ int cps; /* correction per invocation of drbd_rs_controller() */ int steps; /* Number of time steps to plan ahead */ int curr_corr; int max_sect; struct fifo_buffer *plan; - sect_in = atomic_xchg(&mdev->rs_sect_in, 0); /* Number of sectors that came in */ - mdev->rs_in_flight -= sect_in; - - dc = rcu_dereference(mdev->ldev->disk_conf); - plan = rcu_dereference(mdev->rs_plan_s); + dc = rcu_dereference(device->ldev->disk_conf); + plan = rcu_dereference(device->rs_plan_s); steps = plan->size; /* (dc->c_plan_ahead * 10 * SLEEP_TIME) / HZ; */ - if (mdev->rs_in_flight + sect_in == 0) { /* At start of resync */ + if (device->rs_in_flight + sect_in == 0) { /* At start of resync */ want = ((dc->resync_rate * 2 * SLEEP_TIME) / HZ) * steps; } else { /* normal path */ want = dc->c_fill_target ? dc->c_fill_target : sect_in * dc->c_delay_target * HZ / (SLEEP_TIME * 10); } - correction = want - mdev->rs_in_flight - plan->total; + correction = want - device->rs_in_flight - plan->total; /* Plan ahead */ cps = correction / steps; @@ -533,102 +518,123 @@ req_sect = max_sect; /* - dev_warn(DEV, "si=%u if=%d wa=%u co=%d st=%d cps=%d pl=%d cc=%d rs=%d\n", - sect_in, mdev->rs_in_flight, want, correction, - steps, cps, mdev->rs_planed, curr_corr, req_sect); + drbd_warn(device, "si=%u if=%d wa=%u co=%d st=%d cps=%d pl=%d cc=%d rs=%d\n", + sect_in, device->rs_in_flight, want, correction, + steps, cps, device->rs_planed, curr_corr, req_sect); */ return req_sect; } -static int drbd_rs_number_requests(struct drbd_conf *mdev) +static int drbd_rs_number_requests(struct drbd_device *device) { - int number; + unsigned int sect_in; /* Number of sectors that came in since the last turn */ + int number, mxb; + + sect_in = atomic_xchg(&device->rs_sect_in, 0); + device->rs_in_flight -= sect_in; rcu_read_lock(); - if (rcu_dereference(mdev->rs_plan_s)->size) { - number = drbd_rs_controller(mdev) >> (BM_BLOCK_SHIFT - 9); - mdev->c_sync_rate = number * HZ * (BM_BLOCK_SIZE / 1024) / SLEEP_TIME; + mxb = drbd_get_max_buffers(device) / 2; + if (rcu_dereference(device->rs_plan_s)->size) { + number = drbd_rs_controller(device, sect_in) >> (BM_BLOCK_SHIFT - 9); + device->c_sync_rate = number * HZ * (BM_BLOCK_SIZE / 1024) / SLEEP_TIME; } else { - mdev->c_sync_rate = rcu_dereference(mdev->ldev->disk_conf)->resync_rate; - number = SLEEP_TIME * mdev->c_sync_rate / ((BM_BLOCK_SIZE / 1024) * HZ); + device->c_sync_rate = rcu_dereference(device->ldev->disk_conf)->resync_rate; + number = SLEEP_TIME * device->c_sync_rate / ((BM_BLOCK_SIZE / 1024) * HZ); } rcu_read_unlock(); - /* ignore the amount of pending requests, the resync controller should - * throttle down to incoming reply rate soon enough anyways. */ + /* Don't have more than "max-buffers"/2 in-flight. + * Otherwise we may cause the remote site to stall on drbd_alloc_pages(), + * potentially causing a distributed deadlock on congestion during + * online-verify or (checksum-based) resync, if max-buffers, + * socket buffer sizes and resync rate settings are mis-configured. */ + + /* note that "number" is in units of "BM_BLOCK_SIZE" (which is 4k), + * mxb (as used here, and in drbd_alloc_pages on the peer) is + * "number of pages" (typically also 4k), + * but "rs_in_flight" is in "sectors" (512 Byte). */ + if (mxb - device->rs_in_flight/8 < number) + number = mxb - device->rs_in_flight/8; + return number; } -int w_make_resync_request(struct drbd_work *w, int cancel) +static int make_resync_request(struct drbd_device *const device, int cancel) { - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *const peer_device = first_peer_device(device); + struct drbd_connection *const connection = peer_device ? peer_device->connection : NULL; unsigned long bit; sector_t sector; - const sector_t capacity = drbd_get_capacity(mdev->this_bdev); + const sector_t capacity = drbd_get_capacity(device->this_bdev); int max_bio_size; int number, rollback_i, size; - int align, queued, sndbuf; + int align, requeue = 0; int i = 0; if (unlikely(cancel)) return 0; - if (mdev->rs_total == 0) { + if (device->rs_total == 0) { /* empty resync? */ - drbd_resync_finished(mdev); + drbd_resync_finished(device); return 0; } - if (!get_ldev(mdev)) { - /* Since we only need to access mdev->rsync a - get_ldev_if_state(mdev,D_FAILED) would be sufficient, but + if (!get_ldev(device)) { + /* Since we only need to access device->rsync a + get_ldev_if_state(device,D_FAILED) would be sufficient, but to continue resync with a broken disk makes no sense at all */ - dev_err(DEV, "Disk broke down during resync!\n"); + drbd_err(device, "Disk broke down during resync!\n"); return 0; } - max_bio_size = queue_max_hw_sectors(mdev->rq_queue) << 9; - number = drbd_rs_number_requests(mdev); - if (number == 0) + max_bio_size = queue_max_hw_sectors(device->rq_queue) << 9; + number = drbd_rs_number_requests(device); + if (number <= 0) goto requeue; for (i = 0; i < number; i++) { - /* Stop generating RS requests, when half of the send buffer is filled */ - mutex_lock(&mdev->tconn->data.mutex); - if (mdev->tconn->data.socket) { - queued = mdev->tconn->data.socket->sk->sk_wmem_queued; - sndbuf = mdev->tconn->data.socket->sk->sk_sndbuf; - } else { - queued = 1; - sndbuf = 0; - } - mutex_unlock(&mdev->tconn->data.mutex); - if (queued > sndbuf / 2) + /* Stop generating RS requests when half of the send buffer is filled, + * but notify TCP that we'd like to have more space. */ + mutex_lock(&connection->data.mutex); + if (connection->data.socket) { + struct sock *sk = connection->data.socket->sk; + int queued = sk->sk_wmem_queued; + int sndbuf = sk->sk_sndbuf; + if (queued > sndbuf / 2) { + requeue = 1; + if (sk->sk_socket) + set_bit(SOCK_NOSPACE, &sk->sk_socket->flags); + } + } else + requeue = 1; + mutex_unlock(&connection->data.mutex); + if (requeue) goto requeue; next_sector: size = BM_BLOCK_SIZE; - bit = drbd_bm_find_next(mdev, mdev->bm_resync_fo); + bit = drbd_bm_find_next(device, device->bm_resync_fo); if (bit == DRBD_END_OF_BITMAP) { - mdev->bm_resync_fo = drbd_bm_bits(mdev); - put_ldev(mdev); + device->bm_resync_fo = drbd_bm_bits(device); + put_ldev(device); return 0; } sector = BM_BIT_TO_SECT(bit); - if (drbd_rs_should_slow_down(mdev, sector) || - drbd_try_rs_begin_io(mdev, sector)) { - mdev->bm_resync_fo = bit; + if (drbd_try_rs_begin_io(device, sector)) { + device->bm_resync_fo = bit; goto requeue; } - mdev->bm_resync_fo = bit + 1; + device->bm_resync_fo = bit + 1; - if (unlikely(drbd_bm_test_bit(mdev, bit) == 0)) { - drbd_rs_complete_io(mdev, sector); + if (unlikely(drbd_bm_test_bit(device, bit) == 0)) { + drbd_rs_complete_io(device, sector); goto next_sector; } @@ -641,7 +647,7 @@ */ align = 1; rollback_i = i; - for (;;) { + while (i < number) { if (size + BM_BLOCK_SIZE > max_bio_size) break; @@ -657,7 +663,7 @@ * obscure reason; ( b == 0 ) would get the out-of-band * only accidentally right because of the "oddly sized" * adjustment below */ - if (drbd_bm_test_bit(mdev, bit+1) != 1) + if (drbd_bm_test_bit(device, bit+1) != 1) break; bit++; size += BM_BLOCK_SIZE; @@ -668,20 +674,21 @@ /* if we merged some, * reset the offset to start the next drbd_bm_find_next from */ if (size > BM_BLOCK_SIZE) - mdev->bm_resync_fo = bit + 1; + device->bm_resync_fo = bit + 1; #endif /* adjust very last sectors, in case we are oddly sized */ if (sector + (size>>9) > capacity) size = (capacity-sector)<<9; - if (mdev->tconn->agreed_pro_version >= 89 && mdev->tconn->csums_tfm) { - switch (read_for_csum(mdev, sector, size)) { + + if (device->use_csums) { + switch (read_for_csum(peer_device, sector, size)) { case -EIO: /* Disk failure */ - put_ldev(mdev); + put_ldev(device); return -EIO; case -EAGAIN: /* allocation failed, or ldev busy */ - drbd_rs_complete_io(mdev, sector); - mdev->bm_resync_fo = BM_SECT_TO_BIT(sector); + drbd_rs_complete_io(device, sector); + device->bm_resync_fo = BM_SECT_TO_BIT(sector); i = rollback_i; goto requeue; case 0: @@ -693,50 +700,49 @@ } else { int err; - inc_rs_pending(mdev); - err = drbd_send_drequest(mdev, P_RS_DATA_REQUEST, + inc_rs_pending(device); + err = drbd_send_drequest(peer_device, P_RS_DATA_REQUEST, sector, size, ID_SYNCER); if (err) { - dev_err(DEV, "drbd_send_drequest() failed, aborting...\n"); - dec_rs_pending(mdev); - put_ldev(mdev); + drbd_err(device, "drbd_send_drequest() failed, aborting...\n"); + dec_rs_pending(device); + put_ldev(device); return err; } } } - if (mdev->bm_resync_fo >= drbd_bm_bits(mdev)) { + if (device->bm_resync_fo >= drbd_bm_bits(device)) { /* last syncer _request_ was sent, * but the P_RS_DATA_REPLY not yet received. sync will end (and * next sync group will resume), as soon as we receive the last * resync data block, and the last bit is cleared. * until then resync "work" is "inactive" ... */ - put_ldev(mdev); + put_ldev(device); return 0; } requeue: - mdev->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); - mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME); - put_ldev(mdev); + device->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); + mod_timer(&device->resync_timer, jiffies + SLEEP_TIME); + put_ldev(device); return 0; } -static int w_make_ov_request(struct drbd_work *w, int cancel) +static int make_ov_request(struct drbd_device *device, int cancel) { - struct drbd_conf *mdev = w->mdev; int number, i, size; sector_t sector; - const sector_t capacity = drbd_get_capacity(mdev->this_bdev); + const sector_t capacity = drbd_get_capacity(device->this_bdev); bool stop_sector_reached = false; if (unlikely(cancel)) return 1; - number = drbd_rs_number_requests(mdev); + number = drbd_rs_number_requests(device); - sector = mdev->ov_position; + sector = device->ov_position; for (i = 0; i < number; i++) { if (sector >= capacity) return 1; @@ -745,116 +751,120 @@ * w_e_end_ov_reply(). * We need to send at least one request out. */ stop_sector_reached = i > 0 - && verify_can_do_stop_sector(mdev) - && sector >= mdev->ov_stop_sector; + && verify_can_do_stop_sector(device) + && sector >= device->ov_stop_sector; if (stop_sector_reached) break; size = BM_BLOCK_SIZE; - if (drbd_rs_should_slow_down(mdev, sector) || - drbd_try_rs_begin_io(mdev, sector)) { - mdev->ov_position = sector; + if (drbd_try_rs_begin_io(device, sector)) { + device->ov_position = sector; goto requeue; } if (sector + (size>>9) > capacity) size = (capacity-sector)<<9; - inc_rs_pending(mdev); - if (drbd_send_ov_request(mdev, sector, size)) { - dec_rs_pending(mdev); + inc_rs_pending(device); + if (drbd_send_ov_request(first_peer_device(device), sector, size)) { + dec_rs_pending(device); return 0; } sector += BM_SECT_PER_BIT; } - mdev->ov_position = sector; + device->ov_position = sector; requeue: - mdev->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); + device->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); if (i == 0 || !stop_sector_reached) - mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME); + mod_timer(&device->resync_timer, jiffies + SLEEP_TIME); return 1; } int w_ov_finished(struct drbd_work *w, int cancel) { - struct drbd_conf *mdev = w->mdev; - kfree(w); - ov_out_of_sync_print(mdev); - drbd_resync_finished(mdev); + struct drbd_device_work *dw = + container_of(w, struct drbd_device_work, w); + struct drbd_device *device = dw->device; + kfree(dw); + ov_out_of_sync_print(device); + drbd_resync_finished(device); return 0; } static int w_resync_finished(struct drbd_work *w, int cancel) { - struct drbd_conf *mdev = w->mdev; - kfree(w); + struct drbd_device_work *dw = + container_of(w, struct drbd_device_work, w); + struct drbd_device *device = dw->device; + kfree(dw); - drbd_resync_finished(mdev); + drbd_resync_finished(device); return 0; } -static void ping_peer(struct drbd_conf *mdev) +static void ping_peer(struct drbd_device *device) { - struct drbd_tconn *tconn = mdev->tconn; + struct drbd_connection *connection = first_peer_device(device)->connection; - clear_bit(GOT_PING_ACK, &tconn->flags); - request_ping(tconn); - wait_event(tconn->ping_wait, - test_bit(GOT_PING_ACK, &tconn->flags) || mdev->state.conn < C_CONNECTED); + clear_bit(GOT_PING_ACK, &connection->flags); + request_ping(connection); + wait_event(connection->ping_wait, + test_bit(GOT_PING_ACK, &connection->flags) || device->state.conn < C_CONNECTED); } -int drbd_resync_finished(struct drbd_conf *mdev) +int drbd_resync_finished(struct drbd_device *device) { unsigned long db, dt, dbdt; unsigned long n_oos; union drbd_state os, ns; - struct drbd_work *w; + struct drbd_device_work *dw; char *khelper_cmd = NULL; int verify_done = 0; /* Remove all elements from the resync LRU. Since future actions * might set bits in the (main) bitmap, then the entries in the * resync LRU would be wrong. */ - if (drbd_rs_del_all(mdev)) { + if (drbd_rs_del_all(device)) { /* In case this is not possible now, most probably because * there are P_RS_DATA_REPLY Packets lingering on the worker's * queue (or even the read operations for those packets * is not finished by now). Retry in 100ms. */ schedule_timeout_interruptible(HZ / 10); - w = kmalloc(sizeof(struct drbd_work), GFP_ATOMIC); - if (w) { - w->cb = w_resync_finished; - w->mdev = mdev; - drbd_queue_work(&mdev->tconn->sender_work, w); + dw = kmalloc(sizeof(struct drbd_device_work), GFP_ATOMIC); + if (dw) { + dw->w.cb = w_resync_finished; + dw->device = device; + drbd_queue_work(&first_peer_device(device)->connection->sender_work, + &dw->w); return 1; } - dev_err(DEV, "Warn failed to drbd_rs_del_all() and to kmalloc(w).\n"); + drbd_err(device, "Warn failed to drbd_rs_del_all() and to kmalloc(dw).\n"); } - dt = (jiffies - mdev->rs_start - mdev->rs_paused) / HZ; + dt = (jiffies - device->rs_start - device->rs_paused) / HZ; if (dt <= 0) dt = 1; - - db = mdev->rs_total; + + db = device->rs_total; /* adjust for verify start and stop sectors, respective reached position */ - if (mdev->state.conn == C_VERIFY_S || mdev->state.conn == C_VERIFY_T) - db -= mdev->ov_left; + if (device->state.conn == C_VERIFY_S || device->state.conn == C_VERIFY_T) + db -= device->ov_left; dbdt = Bit2KB(db/dt); - mdev->rs_paused /= HZ; + device->rs_paused /= HZ; - if (!get_ldev(mdev)) + if (!get_ldev(device)) goto out; - ping_peer(mdev); + ping_peer(device); - spin_lock_irq(&mdev->tconn->req_lock); - os = drbd_read_state(mdev); + spin_lock_irq(&device->resource->req_lock); + os = drbd_read_state(device); verify_done = (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T); @@ -866,41 +876,41 @@ ns = os; ns.conn = C_CONNECTED; - dev_info(DEV, "%s done (total %lu sec; paused %lu sec; %lu K/sec)\n", + drbd_info(device, "%s done (total %lu sec; paused %lu sec; %lu K/sec)\n", verify_done ? "Online verify" : "Resync", - dt + mdev->rs_paused, mdev->rs_paused, dbdt); + dt + device->rs_paused, device->rs_paused, dbdt); - n_oos = drbd_bm_total_weight(mdev); + n_oos = drbd_bm_total_weight(device); if (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T) { if (n_oos) { - dev_alert(DEV, "Online verify found %lu %dk block out of sync!\n", + drbd_alert(device, "Online verify found %lu %dk block out of sync!\n", n_oos, Bit2KB(1)); khelper_cmd = "out-of-sync"; } } else { - D_ASSERT((n_oos - mdev->rs_failed) == 0); + D_ASSERT(device, (n_oos - device->rs_failed) == 0); if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) khelper_cmd = "after-resync-target"; - if (mdev->tconn->csums_tfm && mdev->rs_total) { - const unsigned long s = mdev->rs_same_csum; - const unsigned long t = mdev->rs_total; + if (device->use_csums && device->rs_total) { + const unsigned long s = device->rs_same_csum; + const unsigned long t = device->rs_total; const int ratio = (t == 0) ? 0 : (t < 100000) ? ((s*100)/t) : (s/(t/100)); - dev_info(DEV, "%u %% had equal checksums, eliminated: %luK; " + drbd_info(device, "%u %% had equal checksums, eliminated: %luK; " "transferred %luK total %luK\n", ratio, - Bit2KB(mdev->rs_same_csum), - Bit2KB(mdev->rs_total - mdev->rs_same_csum), - Bit2KB(mdev->rs_total)); + Bit2KB(device->rs_same_csum), + Bit2KB(device->rs_total - device->rs_same_csum), + Bit2KB(device->rs_total)); } } - if (mdev->rs_failed) { - dev_info(DEV, " %lu failed blocks\n", mdev->rs_failed); + if (device->rs_failed) { + drbd_info(device, " %lu failed blocks\n", device->rs_failed); if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) { ns.disk = D_INCONSISTENT; @@ -914,179 +924,181 @@ ns.pdsk = D_UP_TO_DATE; if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) { - if (mdev->p_uuid) { + if (device->p_uuid) { int i; for (i = UI_BITMAP ; i <= UI_HISTORY_END ; i++) - _drbd_uuid_set(mdev, i, mdev->p_uuid[i]); - drbd_uuid_set(mdev, UI_BITMAP, mdev->ldev->md.uuid[UI_CURRENT]); - _drbd_uuid_set(mdev, UI_CURRENT, mdev->p_uuid[UI_CURRENT]); + _drbd_uuid_set(device, i, device->p_uuid[i]); + drbd_uuid_set(device, UI_BITMAP, device->ldev->md.uuid[UI_CURRENT]); + _drbd_uuid_set(device, UI_CURRENT, device->p_uuid[UI_CURRENT]); } else { - dev_err(DEV, "mdev->p_uuid is NULL! BUG\n"); + drbd_err(device, "device->p_uuid is NULL! BUG\n"); } } if (!(os.conn == C_VERIFY_S || os.conn == C_VERIFY_T)) { /* for verify runs, we don't update uuids here, * so there would be nothing to report. */ - drbd_uuid_set_bm(mdev, 0UL); - drbd_print_uuids(mdev, "updated UUIDs"); - if (mdev->p_uuid) { + drbd_uuid_set_bm(device, 0UL); + drbd_print_uuids(device, "updated UUIDs"); + if (device->p_uuid) { /* Now the two UUID sets are equal, update what we * know of the peer. */ int i; for (i = UI_CURRENT ; i <= UI_HISTORY_END ; i++) - mdev->p_uuid[i] = mdev->ldev->md.uuid[i]; + device->p_uuid[i] = device->ldev->md.uuid[i]; } } } - _drbd_set_state(mdev, ns, CS_VERBOSE, NULL); + _drbd_set_state(device, ns, CS_VERBOSE, NULL); out_unlock: - spin_unlock_irq(&mdev->tconn->req_lock); - put_ldev(mdev); + spin_unlock_irq(&device->resource->req_lock); + put_ldev(device); out: - mdev->rs_total = 0; - mdev->rs_failed = 0; - mdev->rs_paused = 0; + device->rs_total = 0; + device->rs_failed = 0; + device->rs_paused = 0; /* reset start sector, if we reached end of device */ - if (verify_done && mdev->ov_left == 0) - mdev->ov_start_sector = 0; + if (verify_done && device->ov_left == 0) + device->ov_start_sector = 0; - drbd_md_sync(mdev); + drbd_md_sync(device); if (khelper_cmd) - drbd_khelper(mdev, khelper_cmd); + drbd_khelper(device, khelper_cmd); return 1; } /* helper */ -static void move_to_net_ee_or_free(struct drbd_conf *mdev, struct drbd_peer_request *peer_req) +static void move_to_net_ee_or_free(struct drbd_device *device, struct drbd_peer_request *peer_req) { if (drbd_peer_req_has_active_page(peer_req)) { /* This might happen if sendpage() has not finished */ int i = (peer_req->i.size + PAGE_SIZE -1) >> PAGE_SHIFT; - atomic_add(i, &mdev->pp_in_use_by_net); - atomic_sub(i, &mdev->pp_in_use); - spin_lock_irq(&mdev->tconn->req_lock); - list_add_tail(&peer_req->w.list, &mdev->net_ee); - spin_unlock_irq(&mdev->tconn->req_lock); + atomic_add(i, &device->pp_in_use_by_net); + atomic_sub(i, &device->pp_in_use); + spin_lock_irq(&device->resource->req_lock); + list_add_tail(&peer_req->w.list, &device->net_ee); + spin_unlock_irq(&device->resource->req_lock); wake_up(&drbd_pp_wait); } else - drbd_free_peer_req(mdev, peer_req); + drbd_free_peer_req(device, peer_req); } /** * w_e_end_data_req() - Worker callback, to send a P_DATA_REPLY packet in response to a P_DATA_REQUEST - * @mdev: DRBD device. + * @device: DRBD device. * @w: work object. * @cancel: The connection will be closed anyways */ int w_e_end_data_req(struct drbd_work *w, int cancel) { struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; int err; if (unlikely(cancel)) { - drbd_free_peer_req(mdev, peer_req); - dec_unacked(mdev); + drbd_free_peer_req(device, peer_req); + dec_unacked(device); return 0; } if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { - err = drbd_send_block(mdev, P_DATA_REPLY, peer_req); + err = drbd_send_block(peer_device, P_DATA_REPLY, peer_req); } else { if (__ratelimit(&drbd_ratelimit_state)) - dev_err(DEV, "Sending NegDReply. sector=%llus.\n", + drbd_err(device, "Sending NegDReply. sector=%llus.\n", (unsigned long long)peer_req->i.sector); - err = drbd_send_ack(mdev, P_NEG_DREPLY, peer_req); + err = drbd_send_ack(peer_device, P_NEG_DREPLY, peer_req); } - dec_unacked(mdev); + dec_unacked(device); - move_to_net_ee_or_free(mdev, peer_req); + move_to_net_ee_or_free(device, peer_req); if (unlikely(err)) - dev_err(DEV, "drbd_send_block() failed\n"); + drbd_err(device, "drbd_send_block() failed\n"); return err; } /** * w_e_end_rsdata_req() - Worker callback to send a P_RS_DATA_REPLY packet in response to a P_RS_DATA_REQUEST - * @mdev: DRBD device. * @w: work object. * @cancel: The connection will be closed anyways */ int w_e_end_rsdata_req(struct drbd_work *w, int cancel) { struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; int err; if (unlikely(cancel)) { - drbd_free_peer_req(mdev, peer_req); - dec_unacked(mdev); + drbd_free_peer_req(device, peer_req); + dec_unacked(device); return 0; } - if (get_ldev_if_state(mdev, D_FAILED)) { - drbd_rs_complete_io(mdev, peer_req->i.sector); - put_ldev(mdev); + if (get_ldev_if_state(device, D_FAILED)) { + drbd_rs_complete_io(device, peer_req->i.sector); + put_ldev(device); } - if (mdev->state.conn == C_AHEAD) { - err = drbd_send_ack(mdev, P_RS_CANCEL, peer_req); + if (device->state.conn == C_AHEAD) { + err = drbd_send_ack(peer_device, P_RS_CANCEL, peer_req); } else if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { - if (likely(mdev->state.pdsk >= D_INCONSISTENT)) { - inc_rs_pending(mdev); - err = drbd_send_block(mdev, P_RS_DATA_REPLY, peer_req); + if (likely(device->state.pdsk >= D_INCONSISTENT)) { + inc_rs_pending(device); + err = drbd_send_block(peer_device, P_RS_DATA_REPLY, peer_req); } else { if (__ratelimit(&drbd_ratelimit_state)) - dev_err(DEV, "Not sending RSDataReply, " + drbd_err(device, "Not sending RSDataReply, " "partner DISKLESS!\n"); err = 0; } } else { if (__ratelimit(&drbd_ratelimit_state)) - dev_err(DEV, "Sending NegRSDReply. sector %llus.\n", + drbd_err(device, "Sending NegRSDReply. sector %llus.\n", (unsigned long long)peer_req->i.sector); - err = drbd_send_ack(mdev, P_NEG_RS_DREPLY, peer_req); + err = drbd_send_ack(peer_device, P_NEG_RS_DREPLY, peer_req); /* update resync data with failure */ - drbd_rs_failed_io(mdev, peer_req->i.sector, peer_req->i.size); + drbd_rs_failed_io(device, peer_req->i.sector, peer_req->i.size); } - dec_unacked(mdev); + dec_unacked(device); - move_to_net_ee_or_free(mdev, peer_req); + move_to_net_ee_or_free(device, peer_req); if (unlikely(err)) - dev_err(DEV, "drbd_send_block() failed\n"); + drbd_err(device, "drbd_send_block() failed\n"); return err; } int w_e_end_csum_rs_req(struct drbd_work *w, int cancel) { struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; struct digest_info *di; int digest_size; void *digest = NULL; int err, eq = 0; if (unlikely(cancel)) { - drbd_free_peer_req(mdev, peer_req); - dec_unacked(mdev); + drbd_free_peer_req(device, peer_req); + dec_unacked(device); return 0; } - if (get_ldev(mdev)) { - drbd_rs_complete_io(mdev, peer_req->i.sector); - put_ldev(mdev); + if (get_ldev(device)) { + drbd_rs_complete_io(device, peer_req->i.sector); + put_ldev(device); } di = peer_req->digest; @@ -1095,47 +1107,48 @@ /* quick hack to try to avoid a race against reconfiguration. * a real fix would be much more involved, * introducing more locking mechanisms */ - if (mdev->tconn->csums_tfm) { - digest_size = crypto_hash_digestsize(mdev->tconn->csums_tfm); - D_ASSERT(digest_size == di->digest_size); + if (peer_device->connection->csums_tfm) { + digest_size = crypto_hash_digestsize(peer_device->connection->csums_tfm); + D_ASSERT(device, digest_size == di->digest_size); digest = kmalloc(digest_size, GFP_NOIO); } if (digest) { - drbd_csum_ee(mdev, mdev->tconn->csums_tfm, peer_req, digest); + drbd_csum_ee(peer_device->connection->csums_tfm, peer_req, digest); eq = !memcmp(digest, di->digest, digest_size); kfree(digest); } if (eq) { - drbd_set_in_sync(mdev, peer_req->i.sector, peer_req->i.size); + drbd_set_in_sync(device, peer_req->i.sector, peer_req->i.size); /* rs_same_csums unit is BM_BLOCK_SIZE */ - mdev->rs_same_csum += peer_req->i.size >> BM_BLOCK_SHIFT; - err = drbd_send_ack(mdev, P_RS_IS_IN_SYNC, peer_req); + device->rs_same_csum += peer_req->i.size >> BM_BLOCK_SHIFT; + err = drbd_send_ack(peer_device, P_RS_IS_IN_SYNC, peer_req); } else { - inc_rs_pending(mdev); + inc_rs_pending(device); peer_req->block_id = ID_SYNCER; /* By setting block_id, digest pointer becomes invalid! */ peer_req->flags &= ~EE_HAS_DIGEST; /* This peer request no longer has a digest pointer */ kfree(di); - err = drbd_send_block(mdev, P_RS_DATA_REPLY, peer_req); + err = drbd_send_block(peer_device, P_RS_DATA_REPLY, peer_req); } } else { - err = drbd_send_ack(mdev, P_NEG_RS_DREPLY, peer_req); + err = drbd_send_ack(peer_device, P_NEG_RS_DREPLY, peer_req); if (__ratelimit(&drbd_ratelimit_state)) - dev_err(DEV, "Sending NegDReply. I guess it gets messy.\n"); + drbd_err(device, "Sending NegDReply. I guess it gets messy.\n"); } - dec_unacked(mdev); - move_to_net_ee_or_free(mdev, peer_req); + dec_unacked(device); + move_to_net_ee_or_free(device, peer_req); if (unlikely(err)) - dev_err(DEV, "drbd_send_block/ack() failed\n"); + drbd_err(device, "drbd_send_block/ack() failed\n"); return err; } int w_e_end_ov_req(struct drbd_work *w, int cancel) { struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; sector_t sector = peer_req->i.sector; unsigned int size = peer_req->i.size; int digest_size; @@ -1145,7 +1158,7 @@ if (unlikely(cancel)) goto out; - digest_size = crypto_hash_digestsize(mdev->tconn->verify_tfm); + digest_size = crypto_hash_digestsize(peer_device->connection->verify_tfm); digest = kmalloc(digest_size, GFP_NOIO); if (!digest) { err = 1; /* terminate the connection in case the allocation failed */ @@ -1153,7 +1166,7 @@ } if (likely(!(peer_req->flags & EE_WAS_ERROR))) - drbd_csum_ee(mdev, mdev->tconn->verify_tfm, peer_req, digest); + drbd_csum_ee(peer_device->connection->verify_tfm, peer_req, digest); else memset(digest, 0, digest_size); @@ -1162,36 +1175,37 @@ * some distributed deadlock, if the other side blocks on * congestion as well, because our receiver blocks in * drbd_alloc_pages due to pp_in_use > max_buffers. */ - drbd_free_peer_req(mdev, peer_req); + drbd_free_peer_req(device, peer_req); peer_req = NULL; - inc_rs_pending(mdev); - err = drbd_send_drequest_csum(mdev, sector, size, digest, digest_size, P_OV_REPLY); + inc_rs_pending(device); + err = drbd_send_drequest_csum(peer_device, sector, size, digest, digest_size, P_OV_REPLY); if (err) - dec_rs_pending(mdev); + dec_rs_pending(device); kfree(digest); out: if (peer_req) - drbd_free_peer_req(mdev, peer_req); - dec_unacked(mdev); + drbd_free_peer_req(device, peer_req); + dec_unacked(device); return err; } -void drbd_ov_out_of_sync_found(struct drbd_conf *mdev, sector_t sector, int size) +void drbd_ov_out_of_sync_found(struct drbd_device *device, sector_t sector, int size) { - if (mdev->ov_last_oos_start + mdev->ov_last_oos_size == sector) { - mdev->ov_last_oos_size += size>>9; + if (device->ov_last_oos_start + device->ov_last_oos_size == sector) { + device->ov_last_oos_size += size>>9; } else { - mdev->ov_last_oos_start = sector; - mdev->ov_last_oos_size = size>>9; + device->ov_last_oos_start = sector; + device->ov_last_oos_size = size>>9; } - drbd_set_out_of_sync(mdev, sector, size); + drbd_set_out_of_sync(device, sector, size); } int w_e_end_ov_reply(struct drbd_work *w, int cancel) { struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_peer_device *peer_device = peer_req->peer_device; + struct drbd_device *device = peer_device->device; struct digest_info *di; void *digest; sector_t sector = peer_req->i.sector; @@ -1201,27 +1215,27 @@ bool stop_sector_reached = false; if (unlikely(cancel)) { - drbd_free_peer_req(mdev, peer_req); - dec_unacked(mdev); + drbd_free_peer_req(device, peer_req); + dec_unacked(device); return 0; } /* after "cancel", because after drbd_disconnect/drbd_rs_cancel_all * the resync lru has been cleaned up already */ - if (get_ldev(mdev)) { - drbd_rs_complete_io(mdev, peer_req->i.sector); - put_ldev(mdev); + if (get_ldev(device)) { + drbd_rs_complete_io(device, peer_req->i.sector); + put_ldev(device); } di = peer_req->digest; if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { - digest_size = crypto_hash_digestsize(mdev->tconn->verify_tfm); + digest_size = crypto_hash_digestsize(peer_device->connection->verify_tfm); digest = kmalloc(digest_size, GFP_NOIO); if (digest) { - drbd_csum_ee(mdev, mdev->tconn->verify_tfm, peer_req, digest); + drbd_csum_ee(peer_device->connection->verify_tfm, peer_req, digest); - D_ASSERT(digest_size == di->digest_size); + D_ASSERT(device, digest_size == di->digest_size); eq = !memcmp(digest, di->digest, digest_size); kfree(digest); } @@ -1232,116 +1246,111 @@ * some distributed deadlock, if the other side blocks on * congestion as well, because our receiver blocks in * drbd_alloc_pages due to pp_in_use > max_buffers. */ - drbd_free_peer_req(mdev, peer_req); + drbd_free_peer_req(device, peer_req); if (!eq) - drbd_ov_out_of_sync_found(mdev, sector, size); + drbd_ov_out_of_sync_found(device, sector, size); else - ov_out_of_sync_print(mdev); + ov_out_of_sync_print(device); - err = drbd_send_ack_ex(mdev, P_OV_RESULT, sector, size, + err = drbd_send_ack_ex(peer_device, P_OV_RESULT, sector, size, eq ? ID_IN_SYNC : ID_OUT_OF_SYNC); - dec_unacked(mdev); + dec_unacked(device); - --mdev->ov_left; + --device->ov_left; /* let's advance progress step marks only for every other megabyte */ - if ((mdev->ov_left & 0x200) == 0x200) - drbd_advance_rs_marks(mdev, mdev->ov_left); + if ((device->ov_left & 0x200) == 0x200) + drbd_advance_rs_marks(device, device->ov_left); - stop_sector_reached = verify_can_do_stop_sector(mdev) && - (sector + (size>>9)) >= mdev->ov_stop_sector; + stop_sector_reached = verify_can_do_stop_sector(device) && + (sector + (size>>9)) >= device->ov_stop_sector; - if (mdev->ov_left == 0 || stop_sector_reached) { - ov_out_of_sync_print(mdev); - drbd_resync_finished(mdev); + if (device->ov_left == 0 || stop_sector_reached) { + ov_out_of_sync_print(device); + drbd_resync_finished(device); } return err; } -int w_prev_work_done(struct drbd_work *w, int cancel) -{ - struct drbd_wq_barrier *b = container_of(w, struct drbd_wq_barrier, w); - - complete(&b->done); - return 0; -} - /* FIXME * We need to track the number of pending barrier acks, * and to be able to wait for them. * See also comment in drbd_adm_attach before drbd_suspend_io. */ -int drbd_send_barrier(struct drbd_tconn *tconn) +static int drbd_send_barrier(struct drbd_connection *connection) { struct p_barrier *p; struct drbd_socket *sock; - sock = &tconn->data; - p = conn_prepare_command(tconn, sock); + sock = &connection->data; + p = conn_prepare_command(connection, sock); if (!p) return -EIO; - p->barrier = tconn->send.current_epoch_nr; + p->barrier = connection->send.current_epoch_nr; p->pad = 0; - tconn->send.current_epoch_writes = 0; + connection->send.current_epoch_writes = 0; - return conn_send_command(tconn, sock, P_BARRIER, sizeof(*p), NULL, 0); + return conn_send_command(connection, sock, P_BARRIER, sizeof(*p), NULL, 0); } int w_send_write_hint(struct drbd_work *w, int cancel) { - struct drbd_conf *mdev = w->mdev; + struct drbd_device *device = + container_of(w, struct drbd_device, unplug_work); struct drbd_socket *sock; if (cancel) return 0; - sock = &mdev->tconn->data; - if (!drbd_prepare_command(mdev, sock)) + sock = &first_peer_device(device)->connection->data; + if (!drbd_prepare_command(first_peer_device(device), sock)) return -EIO; - return drbd_send_command(mdev, sock, P_UNPLUG_REMOTE, 0, NULL, 0); + return drbd_send_command(first_peer_device(device), sock, P_UNPLUG_REMOTE, 0, NULL, 0); } -static void re_init_if_first_write(struct drbd_tconn *tconn, unsigned int epoch) +static void re_init_if_first_write(struct drbd_connection *connection, unsigned int epoch) { - if (!tconn->send.seen_any_write_yet) { - tconn->send.seen_any_write_yet = true; - tconn->send.current_epoch_nr = epoch; - tconn->send.current_epoch_writes = 0; + if (!connection->send.seen_any_write_yet) { + connection->send.seen_any_write_yet = true; + connection->send.current_epoch_nr = epoch; + connection->send.current_epoch_writes = 0; } } -static void maybe_send_barrier(struct drbd_tconn *tconn, unsigned int epoch) +static void maybe_send_barrier(struct drbd_connection *connection, unsigned int epoch) { /* re-init if first write on this connection */ - if (!tconn->send.seen_any_write_yet) + if (!connection->send.seen_any_write_yet) return; - if (tconn->send.current_epoch_nr != epoch) { - if (tconn->send.current_epoch_writes) - drbd_send_barrier(tconn); - tconn->send.current_epoch_nr = epoch; + if (connection->send.current_epoch_nr != epoch) { + if (connection->send.current_epoch_writes) + drbd_send_barrier(connection); + connection->send.current_epoch_nr = epoch; } } int w_send_out_of_sync(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - struct drbd_conf *mdev = w->mdev; - struct drbd_tconn *tconn = mdev->tconn; + struct drbd_device *device = req->device; + struct drbd_peer_device *const peer_device = first_peer_device(device); + struct drbd_connection *const connection = peer_device->connection; int err; if (unlikely(cancel)) { req_mod(req, SEND_CANCELED); return 0; } + req->pre_send_jif = jiffies; - /* this time, no tconn->send.current_epoch_writes++; + /* this time, no connection->send.current_epoch_writes++; * If it was sent, it was the closing barrier for the last * replicated epoch, before we went into AHEAD mode. * No more barriers will be sent, until we leave AHEAD mode again. */ - maybe_send_barrier(tconn, req->epoch); + maybe_send_barrier(connection, req->epoch); - err = drbd_send_out_of_sync(mdev, req); + err = drbd_send_out_of_sync(peer_device, req); req_mod(req, OOS_HANDED_TO_NETWORK); return err; @@ -1349,27 +1358,28 @@ /** * w_send_dblock() - Worker callback to send a P_DATA packet in order to mirror a write request - * @mdev: DRBD device. * @w: work object. * @cancel: The connection will be closed anyways */ int w_send_dblock(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - struct drbd_conf *mdev = w->mdev; - struct drbd_tconn *tconn = mdev->tconn; + struct drbd_device *device = req->device; + struct drbd_peer_device *const peer_device = first_peer_device(device); + struct drbd_connection *connection = peer_device->connection; int err; if (unlikely(cancel)) { req_mod(req, SEND_CANCELED); return 0; } + req->pre_send_jif = jiffies; - re_init_if_first_write(tconn, req->epoch); - maybe_send_barrier(tconn, req->epoch); - tconn->send.current_epoch_writes++; + re_init_if_first_write(connection, req->epoch); + maybe_send_barrier(connection, req->epoch); + connection->send.current_epoch_writes++; - err = drbd_send_dblock(mdev, req); + err = drbd_send_dblock(peer_device, req); req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK); return err; @@ -1377,27 +1387,28 @@ /** * w_send_read_req() - Worker callback to send a read request (P_DATA_REQUEST) packet - * @mdev: DRBD device. * @w: work object. * @cancel: The connection will be closed anyways */ int w_send_read_req(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - struct drbd_conf *mdev = w->mdev; - struct drbd_tconn *tconn = mdev->tconn; + struct drbd_device *device = req->device; + struct drbd_peer_device *const peer_device = first_peer_device(device); + struct drbd_connection *connection = peer_device->connection; int err; if (unlikely(cancel)) { req_mod(req, SEND_CANCELED); return 0; } + req->pre_send_jif = jiffies; /* Even read requests may close a write epoch, * if there was any yet. */ - maybe_send_barrier(tconn, req->epoch); + maybe_send_barrier(connection, req->epoch); - err = drbd_send_drequest(mdev, P_DATA_REQUEST, req->i.sector, req->i.size, + err = drbd_send_drequest(peer_device, P_DATA_REQUEST, req->i.sector, req->i.size, (unsigned long)req); req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK); @@ -1408,21 +1419,21 @@ int w_restart_disk_io(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - struct drbd_conf *mdev = w->mdev; + struct drbd_device *device = req->device; if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG) - drbd_al_begin_io(mdev, &req->i, false); + drbd_al_begin_io(device, &req->i); drbd_req_make_private_bio(req, req->master_bio); - req->private_bio->bi_bdev = mdev->ldev->backing_bdev; + req->private_bio->bi_bdev = device->ldev->backing_bdev; generic_make_request(req->private_bio); return 0; } -static int _drbd_may_sync_now(struct drbd_conf *mdev) +static int _drbd_may_sync_now(struct drbd_device *device) { - struct drbd_conf *odev = mdev; + struct drbd_device *odev = device; int resync_after; while (1) { @@ -1433,7 +1444,7 @@ rcu_read_unlock(); if (resync_after == -1) return 1; - odev = minor_to_mdev(resync_after); + odev = minor_to_device(resync_after); if (!odev) return 1; if ((odev->state.conn >= C_SYNC_SOURCE && @@ -1446,17 +1457,17 @@ /** * _drbd_pause_after() - Pause resync on all devices that may not resync now - * @mdev: DRBD device. + * @device: DRBD device. * * Called from process context only (admin command and after_state_ch). */ -static int _drbd_pause_after(struct drbd_conf *mdev) +static int _drbd_pause_after(struct drbd_device *device) { - struct drbd_conf *odev; + struct drbd_device *odev; int i, rv = 0; rcu_read_lock(); - idr_for_each_entry(&minors, odev, i) { + idr_for_each_entry(&drbd_devices, odev, i) { if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS) continue; if (!_drbd_may_sync_now(odev)) @@ -1470,17 +1481,17 @@ /** * _drbd_resume_next() - Resume resync on all devices that may resync now - * @mdev: DRBD device. + * @device: DRBD device. * * Called from process context only (admin command and worker). */ -static int _drbd_resume_next(struct drbd_conf *mdev) +static int _drbd_resume_next(struct drbd_device *device) { - struct drbd_conf *odev; + struct drbd_device *odev; int i, rv = 0; rcu_read_lock(); - idr_for_each_entry(&minors, odev, i) { + idr_for_each_entry(&drbd_devices, odev, i) { if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS) continue; if (odev->state.aftr_isp) { @@ -1494,24 +1505,24 @@ return rv; } -void resume_next_sg(struct drbd_conf *mdev) +void resume_next_sg(struct drbd_device *device) { write_lock_irq(&global_state_lock); - _drbd_resume_next(mdev); + _drbd_resume_next(device); write_unlock_irq(&global_state_lock); } -void suspend_other_sg(struct drbd_conf *mdev) +void suspend_other_sg(struct drbd_device *device) { write_lock_irq(&global_state_lock); - _drbd_pause_after(mdev); + _drbd_pause_after(device); write_unlock_irq(&global_state_lock); } /* caller must hold global_state_lock */ -enum drbd_ret_code drbd_resync_after_valid(struct drbd_conf *mdev, int o_minor) +enum drbd_ret_code drbd_resync_after_valid(struct drbd_device *device, int o_minor) { - struct drbd_conf *odev; + struct drbd_device *odev; int resync_after; if (o_minor == -1) @@ -1520,9 +1531,9 @@ return ERR_RESYNC_AFTER; /* check for loops */ - odev = minor_to_mdev(o_minor); + odev = minor_to_device(o_minor); while (1) { - if (odev == mdev) + if (odev == device) return ERR_RESYNC_AFTER_CYCLE; /* You are free to depend on diskless, non-existing, @@ -1542,35 +1553,39 @@ return NO_ERROR; /* follow the dependency chain */ - odev = minor_to_mdev(resync_after); + odev = minor_to_device(resync_after); } } /* caller must hold global_state_lock */ -void drbd_resync_after_changed(struct drbd_conf *mdev) +void drbd_resync_after_changed(struct drbd_device *device) { int changes; do { - changes = _drbd_pause_after(mdev); - changes |= _drbd_resume_next(mdev); + changes = _drbd_pause_after(device); + changes |= _drbd_resume_next(device); } while (changes); } -void drbd_rs_controller_reset(struct drbd_conf *mdev) +void drbd_rs_controller_reset(struct drbd_device *device) { + struct gendisk *disk = device->ldev->backing_bdev->bd_contains->bd_disk; struct fifo_buffer *plan; - atomic_set(&mdev->rs_sect_in, 0); - atomic_set(&mdev->rs_sect_ev, 0); - mdev->rs_in_flight = 0; + atomic_set(&device->rs_sect_in, 0); + atomic_set(&device->rs_sect_ev, 0); + device->rs_in_flight = 0; + device->rs_last_events = + (int)part_stat_read(&disk->part0, sectors[0]) + + (int)part_stat_read(&disk->part0, sectors[1]); /* Updating the RCU protected object in place is necessary since this function gets called from atomic context. It is valid since all other updates also lead to an completely empty fifo */ rcu_read_lock(); - plan = rcu_dereference(mdev->rs_plan_s); + plan = rcu_dereference(device->rs_plan_s); plan->total = 0; fifo_set(plan, 0); rcu_read_unlock(); @@ -1578,101 +1593,116 @@ void start_resync_timer_fn(unsigned long data) { - struct drbd_conf *mdev = (struct drbd_conf *) data; - - drbd_queue_work(&mdev->tconn->sender_work, &mdev->start_resync_work); + struct drbd_device *device = (struct drbd_device *) data; + drbd_device_post_work(device, RS_START); } -int w_start_resync(struct drbd_work *w, int cancel) +static void do_start_resync(struct drbd_device *device) { - struct drbd_conf *mdev = w->mdev; - - if (atomic_read(&mdev->unacked_cnt) || atomic_read(&mdev->rs_pending_cnt)) { - dev_warn(DEV, "w_start_resync later...\n"); - mdev->start_resync_timer.expires = jiffies + HZ/10; - add_timer(&mdev->start_resync_timer); - return 0; + if (atomic_read(&device->unacked_cnt) || atomic_read(&device->rs_pending_cnt)) { + drbd_warn(device, "postponing start_resync ...\n"); + device->start_resync_timer.expires = jiffies + HZ/10; + add_timer(&device->start_resync_timer); + return; } - drbd_start_resync(mdev, C_SYNC_SOURCE); - clear_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags); - return 0; + drbd_start_resync(device, C_SYNC_SOURCE); + clear_bit(AHEAD_TO_SYNC_SOURCE, &device->flags); +} + +static bool use_checksum_based_resync(struct drbd_connection *connection, struct drbd_device *device) +{ + bool csums_after_crash_only; + rcu_read_lock(); + csums_after_crash_only = rcu_dereference(connection->net_conf)->csums_after_crash_only; + rcu_read_unlock(); + return connection->agreed_pro_version >= 89 && /* supported? */ + connection->csums_tfm && /* configured? */ + (csums_after_crash_only == 0 /* use for each resync? */ + || test_bit(CRASHED_PRIMARY, &device->flags)); /* or only after Primary crash? */ } /** * drbd_start_resync() - Start the resync process - * @mdev: DRBD device. + * @device: DRBD device. * @side: Either C_SYNC_SOURCE or C_SYNC_TARGET * * This function might bring you directly into one of the * C_PAUSED_SYNC_* states. */ -void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) +void drbd_start_resync(struct drbd_device *device, enum drbd_conns side) { + struct drbd_peer_device *peer_device = first_peer_device(device); + struct drbd_connection *connection = peer_device ? peer_device->connection : NULL; union drbd_state ns; int r; - if (mdev->state.conn >= C_SYNC_SOURCE && mdev->state.conn < C_AHEAD) { - dev_err(DEV, "Resync already running!\n"); + if (device->state.conn >= C_SYNC_SOURCE && device->state.conn < C_AHEAD) { + drbd_err(device, "Resync already running!\n"); return; } - if (!test_bit(B_RS_H_DONE, &mdev->flags)) { + if (!test_bit(B_RS_H_DONE, &device->flags)) { if (side == C_SYNC_TARGET) { /* Since application IO was locked out during C_WF_BITMAP_T and C_WF_SYNC_UUID we are still unmodified. Before going to C_SYNC_TARGET we check that we might make the data inconsistent. */ - r = drbd_khelper(mdev, "before-resync-target"); + r = drbd_khelper(device, "before-resync-target"); r = (r >> 8) & 0xff; if (r > 0) { - dev_info(DEV, "before-resync-target handler returned %d, " + drbd_info(device, "before-resync-target handler returned %d, " "dropping connection.\n", r); - conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + conn_request_state(connection, NS(conn, C_DISCONNECTING), CS_HARD); return; } } else /* C_SYNC_SOURCE */ { - r = drbd_khelper(mdev, "before-resync-source"); + r = drbd_khelper(device, "before-resync-source"); r = (r >> 8) & 0xff; if (r > 0) { if (r == 3) { - dev_info(DEV, "before-resync-source handler returned %d, " + drbd_info(device, "before-resync-source handler returned %d, " "ignoring. Old userland tools?", r); } else { - dev_info(DEV, "before-resync-source handler returned %d, " + drbd_info(device, "before-resync-source handler returned %d, " "dropping connection.\n", r); - conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + conn_request_state(connection, + NS(conn, C_DISCONNECTING), CS_HARD); return; } } } } - if (current == mdev->tconn->worker.task) { + if (current == connection->worker.task) { /* The worker should not sleep waiting for state_mutex, that can take long */ - if (!mutex_trylock(mdev->state_mutex)) { - set_bit(B_RS_H_DONE, &mdev->flags); - mdev->start_resync_timer.expires = jiffies + HZ/5; - add_timer(&mdev->start_resync_timer); + if (!mutex_trylock(device->state_mutex)) { + set_bit(B_RS_H_DONE, &device->flags); + device->start_resync_timer.expires = jiffies + HZ/5; + add_timer(&device->start_resync_timer); return; } } else { - mutex_lock(mdev->state_mutex); + mutex_lock(device->state_mutex); } - clear_bit(B_RS_H_DONE, &mdev->flags); + clear_bit(B_RS_H_DONE, &device->flags); - write_lock_irq(&global_state_lock); + /* req_lock: serialize with drbd_send_and_submit() and others + * global_state_lock: for stable sync-after dependencies */ + spin_lock_irq(&device->resource->req_lock); + write_lock(&global_state_lock); /* Did some connection breakage or IO error race with us? */ - if (mdev->state.conn < C_CONNECTED - || !get_ldev_if_state(mdev, D_NEGOTIATING)) { - write_unlock_irq(&global_state_lock); - mutex_unlock(mdev->state_mutex); + if (device->state.conn < C_CONNECTED + || !get_ldev_if_state(device, D_NEGOTIATING)) { + write_unlock(&global_state_lock); + spin_unlock_irq(&device->resource->req_lock); + mutex_unlock(device->state_mutex); return; } - ns = drbd_read_state(mdev); + ns = drbd_read_state(device); - ns.aftr_isp = !_drbd_may_sync_now(mdev); + ns.aftr_isp = !_drbd_may_sync_now(device); ns.conn = side; @@ -1681,43 +1711,56 @@ else /* side == C_SYNC_SOURCE */ ns.pdsk = D_INCONSISTENT; - r = __drbd_set_state(mdev, ns, CS_VERBOSE, NULL); - ns = drbd_read_state(mdev); + r = __drbd_set_state(device, ns, CS_VERBOSE, NULL); + ns = drbd_read_state(device); if (ns.conn < C_CONNECTED) r = SS_UNKNOWN_ERROR; if (r == SS_SUCCESS) { - unsigned long tw = drbd_bm_total_weight(mdev); + unsigned long tw = drbd_bm_total_weight(device); unsigned long now = jiffies; int i; - mdev->rs_failed = 0; - mdev->rs_paused = 0; - mdev->rs_same_csum = 0; - mdev->rs_last_events = 0; - mdev->rs_last_sect_ev = 0; - mdev->rs_total = tw; - mdev->rs_start = now; + device->rs_failed = 0; + device->rs_paused = 0; + device->rs_same_csum = 0; + device->rs_last_sect_ev = 0; + device->rs_total = tw; + device->rs_start = now; for (i = 0; i < DRBD_SYNC_MARKS; i++) { - mdev->rs_mark_left[i] = tw; - mdev->rs_mark_time[i] = now; + device->rs_mark_left[i] = tw; + device->rs_mark_time[i] = now; } - _drbd_pause_after(mdev); + _drbd_pause_after(device); + /* Forget potentially stale cached per resync extent bit-counts. + * Open coded drbd_rs_cancel_all(device), we already have IRQs + * disabled, and know the disk state is ok. */ + spin_lock(&device->al_lock); + lc_reset(device->resync); + device->resync_locked = 0; + device->resync_wenr = LC_FREE; + spin_unlock(&device->al_lock); } - write_unlock_irq(&global_state_lock); + write_unlock(&global_state_lock); + spin_unlock_irq(&device->resource->req_lock); if (r == SS_SUCCESS) { + wake_up(&device->al_wait); /* for lc_reset() above */ /* reset rs_last_bcast when a resync or verify is started, * to deal with potential jiffies wrap. */ - mdev->rs_last_bcast = jiffies - HZ; + device->rs_last_bcast = jiffies - HZ; - dev_info(DEV, "Began resync as %s (will sync %lu KB [%lu bits set]).\n", + drbd_info(device, "Began resync as %s (will sync %lu KB [%lu bits set]).\n", drbd_conn_str(ns.conn), - (unsigned long) mdev->rs_total << (BM_BLOCK_SHIFT-10), - (unsigned long) mdev->rs_total); - if (side == C_SYNC_TARGET) - mdev->bm_resync_fo = 0; + (unsigned long) device->rs_total << (BM_BLOCK_SHIFT-10), + (unsigned long) device->rs_total); + if (side == C_SYNC_TARGET) { + device->bm_resync_fo = 0; + device->use_csums = use_checksum_based_resync(connection, device); + } else { + device->use_csums = 0; + } /* Since protocol 96, we must serialize drbd_gen_and_send_sync_uuid * with w_send_oos, or the sync target will get confused as to @@ -1726,10 +1769,10 @@ * drbd_resync_finished from here in that case. * We drbd_gen_and_send_sync_uuid here for protocol < 96, * and from after_state_ch otherwise. */ - if (side == C_SYNC_SOURCE && mdev->tconn->agreed_pro_version < 96) - drbd_gen_and_send_sync_uuid(mdev); + if (side == C_SYNC_SOURCE && connection->agreed_pro_version < 96) + drbd_gen_and_send_sync_uuid(peer_device); - if (mdev->tconn->agreed_pro_version < 95 && mdev->rs_total == 0) { + if (connection->agreed_pro_version < 95 && device->rs_total == 0) { /* This still has a race (about when exactly the peers * detect connection loss) that can lead to a full sync * on next handshake. In 8.3.9 we fixed this with explicit @@ -1745,80 +1788,200 @@ int timeo; rcu_read_lock(); - nc = rcu_dereference(mdev->tconn->net_conf); + nc = rcu_dereference(connection->net_conf); timeo = nc->ping_int * HZ + nc->ping_timeo * HZ / 9; rcu_read_unlock(); schedule_timeout_interruptible(timeo); } - drbd_resync_finished(mdev); + drbd_resync_finished(device); } - drbd_rs_controller_reset(mdev); - /* ns.conn may already be != mdev->state.conn, + drbd_rs_controller_reset(device); + /* ns.conn may already be != device->state.conn, * we may have been paused in between, or become paused until * the timer triggers. * No matter, that is handled in resync_timer_fn() */ if (ns.conn == C_SYNC_TARGET) - mod_timer(&mdev->resync_timer, jiffies); + mod_timer(&device->resync_timer, jiffies); - drbd_md_sync(mdev); + drbd_md_sync(device); } - put_ldev(mdev); - mutex_unlock(mdev->state_mutex); + put_ldev(device); + mutex_unlock(device->state_mutex); } -/* If the resource already closed the current epoch, but we did not - * (because we have not yet seen new requests), we should send the - * corresponding barrier now. Must be checked within the same spinlock - * that is used to check for new requests. */ -bool need_to_send_barrier(struct drbd_tconn *connection) +static void update_on_disk_bitmap(struct drbd_device *device, bool resync_done) { - if (!connection->send.seen_any_write_yet) - return false; + struct sib_info sib = { .sib_reason = SIB_SYNC_PROGRESS, }; + device->rs_last_bcast = jiffies; + + if (!get_ldev(device)) + return; + + drbd_bm_write_lazy(device, 0); + if (resync_done && is_sync_state(device->state.conn)) + drbd_resync_finished(device); - /* Skip barriers that do not contain any writes. - * This may happen during AHEAD mode. */ - if (!connection->send.current_epoch_writes) - return false; - - /* ->req_lock is held when requests are queued on - * connection->sender_work, and put into ->transfer_log. - * It is also held when ->current_tle_nr is increased. - * So either there are already new requests queued, - * and corresponding barriers will be send there. - * Or nothing new is queued yet, so the difference will be 1. + drbd_bcast_event(device, &sib); + /* update timestamp, in case it took a while to write out stuff */ + device->rs_last_bcast = jiffies; + put_ldev(device); +} + +static void drbd_ldev_destroy(struct drbd_device *device) +{ + lc_destroy(device->resync); + device->resync = NULL; + lc_destroy(device->act_log); + device->act_log = NULL; + + __acquire(local); + drbd_free_ldev(device->ldev); + device->ldev = NULL; + __release(local); + + clear_bit(GOING_DISKLESS, &device->flags); + wake_up(&device->misc_wait); +} + +static void go_diskless(struct drbd_device *device) +{ + D_ASSERT(device, device->state.disk == D_FAILED); + /* we cannot assert local_cnt == 0 here, as get_ldev_if_state will + * inc/dec it frequently. Once we are D_DISKLESS, no one will touch + * the protected members anymore, though, so once put_ldev reaches zero + * again, it will be safe to free them. */ + + /* Try to write changed bitmap pages, read errors may have just + * set some bits outside the area covered by the activity log. + * + * If we have an IO error during the bitmap writeout, + * we will want a full sync next time, just in case. + * (Do we want a specific meta data flag for this?) + * + * If that does not make it to stable storage either, + * we cannot do anything about that anymore. + * + * We still need to check if both bitmap and ldev are present, we may + * end up here after a failed attach, before ldev was even assigned. */ - if (atomic_read(&connection->current_tle_nr) != - connection->send.current_epoch_nr + 1) - return false; + if (device->bitmap && device->ldev) { + /* An interrupted resync or similar is allowed to recounts bits + * while we detach. + * Any modifications would not be expected anymore, though. + */ + if (drbd_bitmap_io_from_worker(device, drbd_bm_write, + "detach", BM_LOCKED_TEST_ALLOWED)) { + if (test_bit(WAS_READ_ERROR, &device->flags)) { + drbd_md_set_flag(device, MDF_FULL_SYNC); + drbd_md_sync(device); + } + } + } - return true; + drbd_force_state(device, NS(disk, D_DISKLESS)); } -bool dequeue_work_batch(struct drbd_work_queue *queue, struct list_head *work_list) +static int do_md_sync(struct drbd_device *device) { - spin_lock_irq(&queue->q_lock); - list_splice_init(&queue->q, work_list); - spin_unlock_irq(&queue->q_lock); - return !list_empty(work_list); + drbd_warn(device, "md_sync_timer expired! Worker calls drbd_md_sync().\n"); + drbd_md_sync(device); + return 0; } -bool dequeue_work_item(struct drbd_work_queue *queue, struct list_head *work_list) +/* only called from drbd_worker thread, no locking */ +void __update_timing_details( + struct drbd_thread_timing_details *tdp, + unsigned int *cb_nr, + void *cb, + const char *fn, const unsigned int line) +{ + unsigned int i = *cb_nr % DRBD_THREAD_DETAILS_HIST; + struct drbd_thread_timing_details *td = tdp + i; + + td->start_jif = jiffies; + td->cb_addr = cb; + td->caller_fn = fn; + td->line = line; + td->cb_nr = *cb_nr; + + i = (i+1) % DRBD_THREAD_DETAILS_HIST; + td = tdp + i; + memset(td, 0, sizeof(*td)); + + ++(*cb_nr); +} + +static void do_device_work(struct drbd_device *device, const unsigned long todo) +{ + if (test_bit(MD_SYNC, &todo)) + do_md_sync(device); + if (test_bit(RS_DONE, &todo) || + test_bit(RS_PROGRESS, &todo)) + update_on_disk_bitmap(device, test_bit(RS_DONE, &todo)); + if (test_bit(GO_DISKLESS, &todo)) + go_diskless(device); + if (test_bit(DESTROY_DISK, &todo)) + drbd_ldev_destroy(device); + if (test_bit(RS_START, &todo)) + do_start_resync(device); +} + +#define DRBD_DEVICE_WORK_MASK \ + ((1UL << GO_DISKLESS) \ + |(1UL << DESTROY_DISK) \ + |(1UL << MD_SYNC) \ + |(1UL << RS_START) \ + |(1UL << RS_PROGRESS) \ + |(1UL << RS_DONE) \ + ) + +static unsigned long get_work_bits(unsigned long *flags) +{ + unsigned long old, new; + do { + old = *flags; + new = old & ~DRBD_DEVICE_WORK_MASK; + } while (cmpxchg(flags, old, new) != old); + return old & DRBD_DEVICE_WORK_MASK; +} + +static void do_unqueued_work(struct drbd_connection *connection) +{ + struct drbd_peer_device *peer_device; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&connection->peer_devices, peer_device, vnr) { + struct drbd_device *device = peer_device->device; + unsigned long todo = get_work_bits(&device->flags); + if (!todo) + continue; + + kref_get(&device->kref); + rcu_read_unlock(); + do_device_work(device, todo); + kref_put(&device->kref, drbd_destroy_device); + rcu_read_lock(); + } + rcu_read_unlock(); +} + +static bool dequeue_work_batch(struct drbd_work_queue *queue, struct list_head *work_list) { spin_lock_irq(&queue->q_lock); - if (!list_empty(&queue->q)) - list_move(queue->q.next, work_list); + list_splice_tail_init(&queue->q, work_list); spin_unlock_irq(&queue->q_lock); return !list_empty(work_list); } -void wait_for_work(struct drbd_tconn *connection, struct list_head *work_list) +static void wait_for_work(struct drbd_connection *connection, struct list_head *work_list) { DEFINE_WAIT(wait); struct net_conf *nc; int uncork, cork; - dequeue_work_item(&connection->sender_work, work_list); + dequeue_work_batch(&connection->sender_work, work_list); if (!list_empty(work_list)) return; @@ -1842,23 +2005,39 @@ for (;;) { int send_barrier; prepare_to_wait(&connection->sender_work.q_wait, &wait, TASK_INTERRUPTIBLE); - spin_lock_irq(&connection->req_lock); + spin_lock_irq(&connection->resource->req_lock); spin_lock(&connection->sender_work.q_lock); /* FIXME get rid of this one? */ - /* dequeue single item only, - * we still use drbd_queue_work_front() in some places */ if (!list_empty(&connection->sender_work.q)) - list_move(connection->sender_work.q.next, work_list); + list_splice_tail_init(&connection->sender_work.q, work_list); spin_unlock(&connection->sender_work.q_lock); /* FIXME get rid of this one? */ if (!list_empty(work_list) || signal_pending(current)) { - spin_unlock_irq(&connection->req_lock); + spin_unlock_irq(&connection->resource->req_lock); break; } - send_barrier = need_to_send_barrier(connection); - spin_unlock_irq(&connection->req_lock); - if (send_barrier) { - drbd_send_barrier(connection); - connection->send.current_epoch_nr++; - } + + /* We found nothing new to do, no to-be-communicated request, + * no other work item. We may still need to close the last + * epoch. Next incoming request epoch will be connection -> + * current transfer log epoch number. If that is different + * from the epoch of the last request we communicated, it is + * safe to send the epoch separating barrier now. + */ + send_barrier = + atomic_read(&connection->current_tle_nr) != + connection->send.current_epoch_nr; + spin_unlock_irq(&connection->resource->req_lock); + + if (send_barrier) + maybe_send_barrier(connection, + connection->send.current_epoch_nr + 1); + + if (test_bit(DEVICE_WORK_PENDING, &connection->flags)) + break; + + /* drbd_send() may have called flush_signals() */ + if (get_t_state(&connection->worker) != RUNNING) + break; + schedule(); /* may be woken up for other things but new work, too, * e.g. if the current epoch got closed. @@ -1883,24 +2062,29 @@ int drbd_worker(struct drbd_thread *thi) { - struct drbd_tconn *tconn = thi->tconn; + struct drbd_connection *connection = thi->connection; struct drbd_work *w = NULL; - struct drbd_conf *mdev; + struct drbd_peer_device *peer_device; LIST_HEAD(work_list); int vnr; while (get_t_state(thi) == RUNNING) { drbd_thread_current_set_cpu(thi); - /* as long as we use drbd_queue_work_front(), - * we may only dequeue single work items here, not batches. */ - if (list_empty(&work_list)) - wait_for_work(tconn, &work_list); + if (list_empty(&work_list)) { + update_worker_timing_details(connection, wait_for_work); + wait_for_work(connection, &work_list); + } + + if (test_and_clear_bit(DEVICE_WORK_PENDING, &connection->flags)) { + update_worker_timing_details(connection, do_unqueued_work); + do_unqueued_work(connection); + } if (signal_pending(current)) { flush_signals(current); if (get_t_state(thi) == RUNNING) { - conn_warn(tconn, "Worker got an unexpected signal\n"); + drbd_warn(connection, "Worker got an unexpected signal\n"); continue; } break; @@ -1909,32 +2093,39 @@ if (get_t_state(thi) != RUNNING) break; - while (!list_empty(&work_list)) { + if (!list_empty(&work_list)) { w = list_first_entry(&work_list, struct drbd_work, list); list_del_init(&w->list); - if (w->cb(w, tconn->cstate < C_WF_REPORT_PARAMS) == 0) + update_worker_timing_details(connection, w->cb); + if (w->cb(w, connection->cstate < C_WF_REPORT_PARAMS) == 0) continue; - if (tconn->cstate >= C_WF_REPORT_PARAMS) - conn_request_state(tconn, NS(conn, C_NETWORK_FAILURE), CS_HARD); + if (connection->cstate >= C_WF_REPORT_PARAMS) + conn_request_state(connection, NS(conn, C_NETWORK_FAILURE), CS_HARD); } } do { - while (!list_empty(&work_list)) { + if (test_and_clear_bit(DEVICE_WORK_PENDING, &connection->flags)) { + update_worker_timing_details(connection, do_unqueued_work); + do_unqueued_work(connection); + } + if (!list_empty(&work_list)) { w = list_first_entry(&work_list, struct drbd_work, list); list_del_init(&w->list); + update_worker_timing_details(connection, w->cb); w->cb(w, 1); - } - dequeue_work_batch(&tconn->sender_work, &work_list); - } while (!list_empty(&work_list)); + } else + dequeue_work_batch(&connection->sender_work, &work_list); + } while (!list_empty(&work_list) || test_bit(DEVICE_WORK_PENDING, &connection->flags)); rcu_read_lock(); - idr_for_each_entry(&tconn->volumes, mdev, vnr) { - D_ASSERT(mdev->state.disk == D_DISKLESS && mdev->state.conn == C_STANDALONE); - kref_get(&mdev->kref); + idr_for_each_entry(&connection->peer_devices, peer_device, vnr) { + struct drbd_device *device = peer_device->device; + D_ASSERT(device, device->state.disk == D_DISKLESS && device->state.conn == C_STANDALONE); + kref_get(&device->kref); rcu_read_unlock(); - drbd_mdev_cleanup(mdev); - kref_put(&mdev->kref, &drbd_minor_destroy); + drbd_device_cleanup(device); + kref_put(&device->kref, drbd_destroy_device); rcu_read_lock(); } rcu_read_unlock();