summaryrefslogtreecommitdiff
path: root/drivers/block
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/block')
-rw-r--r--drivers/block/drbd/drbd_actlog.c104
-rw-r--r--drivers/block/drbd/drbd_bitmap.c146
-rw-r--r--drivers/block/drbd/drbd_int.h90
-rw-r--r--drivers/block/drbd/drbd_main.c357
-rw-r--r--drivers/block/drbd/drbd_nl.c48
-rw-r--r--drivers/block/drbd/drbd_proc.c2
-rw-r--r--drivers/block/drbd/drbd_receiver.c101
-rw-r--r--drivers/block/drbd/drbd_req.c132
-rw-r--r--drivers/block/drbd/drbd_req.h19
-rw-r--r--drivers/block/drbd/drbd_worker.c31
-rw-r--r--drivers/block/floppy.c161
-rw-r--r--drivers/block/mtip32xx/mtip32xx.c276
-rw-r--r--drivers/block/mtip32xx/mtip32xx.h48
-rw-r--r--drivers/block/rbd.c72
-rw-r--r--drivers/block/virtio_blk.c21
-rw-r--r--drivers/block/xen-blkfront.c44
16 files changed, 1083 insertions, 569 deletions
diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index cf0e63dd97da..e54e31b02b88 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -65,39 +65,80 @@ struct drbd_atodb_wait {
int w_al_write_transaction(struct drbd_conf *, struct drbd_work *, int);
+void *drbd_md_get_buffer(struct drbd_conf *mdev)
+{
+ int r;
+
+ wait_event(mdev->misc_wait,
+ (r = atomic_cmpxchg(&mdev->md_io_in_use, 0, 1)) == 0 ||
+ mdev->state.disk <= D_FAILED);
+
+ return r ? NULL : page_address(mdev->md_io_page);
+}
+
+void drbd_md_put_buffer(struct drbd_conf *mdev)
+{
+ if (atomic_dec_and_test(&mdev->md_io_in_use))
+ wake_up(&mdev->misc_wait);
+}
+
+static bool md_io_allowed(struct drbd_conf *mdev)
+{
+ enum drbd_disk_state ds = mdev->state.disk;
+ return ds >= D_NEGOTIATING || ds == D_ATTACHING;
+}
+
+void wait_until_done_or_disk_failure(struct drbd_conf *mdev, struct drbd_backing_dev *bdev,
+ unsigned int *done)
+{
+ long dt = bdev->dc.disk_timeout * HZ / 10;
+ if (dt == 0)
+ dt = MAX_SCHEDULE_TIMEOUT;
+
+ dt = wait_event_timeout(mdev->misc_wait, *done || !md_io_allowed(mdev), dt);
+ if (dt == 0)
+ dev_err(DEV, "meta-data IO operation timed out\n");
+}
+
static int _drbd_md_sync_page_io(struct drbd_conf *mdev,
struct drbd_backing_dev *bdev,
struct page *page, sector_t sector,
int rw, int size)
{
struct bio *bio;
- struct drbd_md_io md_io;
int ok;
- md_io.mdev = mdev;
- init_completion(&md_io.event);
- md_io.error = 0;
+ mdev->md_io.done = 0;
+ mdev->md_io.error = -ENODEV;
if ((rw & WRITE) && !test_bit(MD_NO_FUA, &mdev->flags))
rw |= REQ_FUA | REQ_FLUSH;
rw |= REQ_SYNC;
- bio = bio_alloc(GFP_NOIO, 1);
+ bio = bio_alloc_drbd(GFP_NOIO);
bio->bi_bdev = bdev->md_bdev;
bio->bi_sector = sector;
ok = (bio_add_page(bio, page, size, 0) == size);
if (!ok)
goto out;
- bio->bi_private = &md_io;
+ bio->bi_private = &mdev->md_io;
bio->bi_end_io = drbd_md_io_complete;
bio->bi_rw = rw;
+ if (!get_ldev_if_state(mdev, D_ATTACHING)) { /* Corresponding put_ldev in drbd_md_io_complete() */
+ dev_err(DEV, "ASSERT FAILED: get_ldev_if_state() == 1 in _drbd_md_sync_page_io()\n");
+ ok = 0;
+ goto out;
+ }
+
+ bio_get(bio); /* one bio_put() is in the completion handler */
+ atomic_inc(&mdev->md_io_in_use); /* drbd_md_put_buffer() is in the completion handler */
if (drbd_insert_fault(mdev, (rw & WRITE) ? DRBD_FAULT_MD_WR : DRBD_FAULT_MD_RD))
bio_endio(bio, -EIO);
else
submit_bio(rw, bio);
- wait_for_completion(&md_io.event);
- ok = bio_flagged(bio, BIO_UPTODATE) && md_io.error == 0;
+ wait_until_done_or_disk_failure(mdev, bdev, &mdev->md_io.done);
+ ok = bio_flagged(bio, BIO_UPTODATE) && mdev->md_io.error == 0;
out:
bio_put(bio);
@@ -111,7 +152,7 @@ int drbd_md_sync_page_io(struct drbd_conf *mdev, struct drbd_backing_dev *bdev,
int offset = 0;
struct page *iop = mdev->md_io_page;
- D_ASSERT(mutex_is_locked(&mdev->md_io_mutex));
+ D_ASSERT(atomic_read(&mdev->md_io_in_use) == 1);
BUG_ON(!bdev->md_bdev);
@@ -328,8 +369,13 @@ w_al_write_transaction(struct drbd_conf *mdev, struct drbd_work *w, int unused)
return 1;
}
- mutex_lock(&mdev->md_io_mutex); /* protects md_io_buffer, al_tr_cycle, ... */
- buffer = (struct al_transaction *)page_address(mdev->md_io_page);
+ buffer = drbd_md_get_buffer(mdev); /* protects md_io_buffer, al_tr_cycle, ... */
+ if (!buffer) {
+ dev_err(DEV, "disk failed while waiting for md_io buffer\n");
+ complete(&((struct update_al_work *)w)->event);
+ put_ldev(mdev);
+ return 1;
+ }
buffer->magic = __constant_cpu_to_be32(DRBD_MAGIC);
buffer->tr_number = cpu_to_be32(mdev->al_tr_number);
@@ -374,7 +420,7 @@ w_al_write_transaction(struct drbd_conf *mdev, struct drbd_work *w, int unused)
D_ASSERT(mdev->al_tr_pos < MD_AL_MAX_SIZE);
mdev->al_tr_number++;
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
complete(&((struct update_al_work *)w)->event);
put_ldev(mdev);
@@ -443,8 +489,9 @@ int drbd_al_read_log(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
/* lock out all other meta data io for now,
* and make sure the page is mapped.
*/
- mutex_lock(&mdev->md_io_mutex);
- buffer = page_address(mdev->md_io_page);
+ buffer = drbd_md_get_buffer(mdev);
+ if (!buffer)
+ return 0;
/* Find the valid transaction in the log */
for (i = 0; i <= mx; i++) {
@@ -452,7 +499,7 @@ int drbd_al_read_log(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
if (rv == 0)
continue;
if (rv == -1) {
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
return 0;
}
cnr = be32_to_cpu(buffer->tr_number);
@@ -478,7 +525,7 @@ int drbd_al_read_log(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
if (!found_valid) {
dev_warn(DEV, "No usable activity log found.\n");
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
return 1;
}
@@ -493,7 +540,7 @@ int drbd_al_read_log(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
rv = drbd_al_read_tr(mdev, bdev, buffer, i);
ERR_IF(rv == 0) goto cancel;
if (rv == -1) {
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
return 0;
}
@@ -534,7 +581,7 @@ cancel:
mdev->al_tr_pos = 0;
/* ok, we are done with it */
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
dev_info(DEV, "Found %d transactions (%d active extents) in activity log.\n",
transactions, active_extents);
@@ -671,16 +718,20 @@ static void drbd_try_clear_on_disk_bm(struct drbd_conf *mdev, sector_t sector,
else
ext->rs_failed += count;
if (ext->rs_left < ext->rs_failed) {
- dev_err(DEV, "BAD! sector=%llus enr=%u rs_left=%d "
- "rs_failed=%d count=%d\n",
+ dev_warn(DEV, "BAD! sector=%llus enr=%u rs_left=%d "
+ "rs_failed=%d count=%d cstate=%s\n",
(unsigned long long)sector,
ext->lce.lc_number, ext->rs_left,
- ext->rs_failed, count);
- dump_stack();
-
- lc_put(mdev->resync, &ext->lce);
- drbd_force_state(mdev, NS(conn, C_DISCONNECTING));
- return;
+ ext->rs_failed, count,
+ drbd_conn_str(mdev->state.conn));
+
+ /* We don't expect to be able to clear more bits
+ * than have been set when we originally counted
+ * the set bits to cache that value in ext->rs_left.
+ * Whatever the reason (disconnect during resync,
+ * delayed local completion of an application write),
+ * try to fix it up by recounting here. */
+ ext->rs_left = drbd_bm_e_weight(mdev, enr);
}
} else {
/* Normally this element should be in the cache,
@@ -1192,6 +1243,7 @@ int drbd_rs_del_all(struct drbd_conf *mdev)
put_ldev(mdev);
}
spin_unlock_irq(&mdev->al_lock);
+ wake_up(&mdev->al_wait);
return 0;
}
diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index 3030201c69d8..b5c5ff53cb57 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -205,7 +205,7 @@ void drbd_bm_unlock(struct drbd_conf *mdev)
static void bm_store_page_idx(struct page *page, unsigned long idx)
{
BUG_ON(0 != (idx & ~BM_PAGE_IDX_MASK));
- page_private(page) |= idx;
+ set_page_private(page, idx);
}
static unsigned long bm_page_to_idx(struct page *page)
@@ -886,12 +886,21 @@ void drbd_bm_clear_all(struct drbd_conf *mdev)
struct bm_aio_ctx {
struct drbd_conf *mdev;
atomic_t in_flight;
- struct completion done;
+ unsigned int done;
unsigned flags;
#define BM_AIO_COPY_PAGES 1
int error;
+ struct kref kref;
};
+static void bm_aio_ctx_destroy(struct kref *kref)
+{
+ struct bm_aio_ctx *ctx = container_of(kref, struct bm_aio_ctx, kref);
+
+ put_ldev(ctx->mdev);
+ kfree(ctx);
+}
+
/* bv_page may be a copy, or may be the original */
static void bm_async_io_complete(struct bio *bio, int error)
{
@@ -930,20 +939,21 @@ static void bm_async_io_complete(struct bio *bio, int error)
bm_page_unlock_io(mdev, idx);
- /* FIXME give back to page pool */
if (ctx->flags & BM_AIO_COPY_PAGES)
- put_page(bio->bi_io_vec[0].bv_page);
+ mempool_free(bio->bi_io_vec[0].bv_page, drbd_md_io_page_pool);
bio_put(bio);
- if (atomic_dec_and_test(&ctx->in_flight))
- complete(&ctx->done);
+ if (atomic_dec_and_test(&ctx->in_flight)) {
+ ctx->done = 1;
+ wake_up(&mdev->misc_wait);
+ kref_put(&ctx->kref, &bm_aio_ctx_destroy);
+ }
}
static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must_hold(local)
{
- /* we are process context. we always get a bio */
- struct bio *bio = bio_alloc(GFP_KERNEL, 1);
+ struct bio *bio = bio_alloc_drbd(GFP_NOIO);
struct drbd_conf *mdev = ctx->mdev;
struct drbd_bitmap *b = mdev->bitmap;
struct page *page;
@@ -966,10 +976,8 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must
bm_set_page_unchanged(b->bm_pages[page_nr]);
if (ctx->flags & BM_AIO_COPY_PAGES) {
- /* FIXME alloc_page is good enough for now, but actually needs
- * to use pre-allocated page pool */
void *src, *dest;
- page = alloc_page(__GFP_HIGHMEM|__GFP_WAIT);
+ page = mempool_alloc(drbd_md_io_page_pool, __GFP_HIGHMEM|__GFP_WAIT);
dest = kmap_atomic(page);
src = kmap_atomic(b->bm_pages[page_nr]);
memcpy(dest, src, PAGE_SIZE);
@@ -981,6 +989,8 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must
bio->bi_bdev = mdev->ldev->md_bdev;
bio->bi_sector = on_disk_sector;
+ /* bio_add_page of a single page to an empty bio will always succeed,
+ * according to api. Do we want to assert that? */
bio_add_page(bio, page, len, 0);
bio->bi_private = ctx;
bio->bi_end_io = bm_async_io_complete;
@@ -999,14 +1009,9 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must
/*
* bm_rw: read/write the whole bitmap from/to its on disk location.
*/
-static int bm_rw(struct drbd_conf *mdev, int rw, unsigned lazy_writeout_upper_idx) __must_hold(local)
+static int bm_rw(struct drbd_conf *mdev, int rw, unsigned flags, unsigned lazy_writeout_upper_idx) __must_hold(local)
{
- struct bm_aio_ctx ctx = {
- .mdev = mdev,
- .in_flight = ATOMIC_INIT(1),
- .done = COMPLETION_INITIALIZER_ONSTACK(ctx.done),
- .flags = lazy_writeout_upper_idx ? BM_AIO_COPY_PAGES : 0,
- };
+ struct bm_aio_ctx *ctx;
struct drbd_bitmap *b = mdev->bitmap;
int num_pages, i, count = 0;
unsigned long now;
@@ -1021,7 +1026,27 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned lazy_writeout_upper_id
* For lazy writeout, we don't care for ongoing changes to the bitmap,
* as we submit copies of pages anyways.
*/
- if (!ctx.flags)
+
+ ctx = kmalloc(sizeof(struct bm_aio_ctx), GFP_NOIO);
+ if (!ctx)
+ return -ENOMEM;
+
+ *ctx = (struct bm_aio_ctx) {
+ .mdev = mdev,
+ .in_flight = ATOMIC_INIT(1),
+ .done = 0,
+ .flags = flags,
+ .error = 0,
+ .kref = { ATOMIC_INIT(2) },
+ };
+
+ if (!get_ldev_if_state(mdev, D_ATTACHING)) { /* put is in bm_aio_ctx_destroy() */
+ dev_err(DEV, "ASSERT FAILED: get_ldev_if_state() == 1 in bm_rw()\n");
+ kfree(ctx);
+ return -ENODEV;
+ }
+
+ if (!ctx->flags)
WARN_ON(!(BM_LOCKED_MASK & b->bm_flags));
num_pages = b->bm_number_of_pages;
@@ -1046,29 +1071,38 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned lazy_writeout_upper_id
continue;
}
}
- atomic_inc(&ctx.in_flight);
- bm_page_io_async(&ctx, i, rw);
+ atomic_inc(&ctx->in_flight);
+ bm_page_io_async(ctx, i, rw);
++count;
cond_resched();
}
/*
- * We initialize ctx.in_flight to one to make sure bm_async_io_complete
- * will not complete() early, and decrement / test it here. If there
+ * We initialize ctx->in_flight to one to make sure bm_async_io_complete
+ * will not set ctx->done early, and decrement / test it here. If there
* are still some bios in flight, we need to wait for them here.
+ * If all IO is done already (or nothing had been submitted), there is
+ * no need to wait. Still, we need to put the kref associated with the
+ * "in_flight reached zero, all done" event.
*/
- if (!atomic_dec_and_test(&ctx.in_flight))
- wait_for_completion(&ctx.done);
+ if (!atomic_dec_and_test(&ctx->in_flight))
+ wait_until_done_or_disk_failure(mdev, mdev->ldev, &ctx->done);
+ else
+ kref_put(&ctx->kref, &bm_aio_ctx_destroy);
+
dev_info(DEV, "bitmap %s of %u pages took %lu jiffies\n",
rw == WRITE ? "WRITE" : "READ",
count, jiffies - now);
- if (ctx.error) {
+ 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);
- err = -EIO; /* ctx.error ? */
+ err = -EIO; /* ctx->error ? */
}
+ if (atomic_read(&ctx->in_flight))
+ err = -EIO; /* Disk failed during IO... */
+
now = jiffies;
if (rw == WRITE) {
drbd_md_flush(mdev);
@@ -1082,6 +1116,7 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned lazy_writeout_upper_id
dev_info(DEV, "%s (%lu bits) marked out-of-sync by on disk bit-map.\n",
ppsize(ppb, now << (BM_BLOCK_SHIFT-10)), now);
+ kref_put(&ctx->kref, &bm_aio_ctx_destroy);
return err;
}
@@ -1091,7 +1126,7 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned lazy_writeout_upper_id
*/
int drbd_bm_read(struct drbd_conf *mdev) __must_hold(local)
{
- return bm_rw(mdev, READ, 0);
+ return bm_rw(mdev, READ, 0, 0);
}
/**
@@ -1102,7 +1137,7 @@ int drbd_bm_read(struct drbd_conf *mdev) __must_hold(local)
*/
int drbd_bm_write(struct drbd_conf *mdev) __must_hold(local)
{
- return bm_rw(mdev, WRITE, 0);
+ return bm_rw(mdev, WRITE, 0, 0);
}
/**
@@ -1112,7 +1147,23 @@ int drbd_bm_write(struct drbd_conf *mdev) __must_hold(local)
*/
int drbd_bm_write_lazy(struct drbd_conf *mdev, unsigned upper_idx) __must_hold(local)
{
- return bm_rw(mdev, WRITE, upper_idx);
+ return bm_rw(mdev, WRITE, BM_AIO_COPY_PAGES, upper_idx);
+}
+
+/**
+ * drbd_bm_write_copy_pages() - Write the whole bitmap to its on disk location.
+ * @mdev: DRBD device.
+ *
+ * Will only write pages that have changed since last IO.
+ * In contrast to drbd_bm_write(), this will copy the bitmap pages
+ * to temporary writeout pages. It is intended to trigger a full write-out
+ * while still allowing the bitmap to change, for example if a resync or online
+ * verify is aborted due to a failed peer disk, while local IO continues, or
+ * pending resync acks are still being processed.
+ */
+int drbd_bm_write_copy_pages(struct drbd_conf *mdev) __must_hold(local)
+{
+ return bm_rw(mdev, WRITE, BM_AIO_COPY_PAGES, 0);
}
@@ -1130,28 +1181,45 @@ int drbd_bm_write_lazy(struct drbd_conf *mdev, unsigned upper_idx) __must_hold(l
*/
int drbd_bm_write_page(struct drbd_conf *mdev, unsigned int idx) __must_hold(local)
{
- struct bm_aio_ctx ctx = {
+ struct bm_aio_ctx *ctx;
+ int err;
+
+ if (bm_test_page_unchanged(mdev->bitmap->bm_pages[idx])) {
+ dynamic_dev_dbg(DEV, "skipped bm page write for idx %u\n", idx);
+ return 0;
+ }
+
+ ctx = kmalloc(sizeof(struct bm_aio_ctx), GFP_NOIO);
+ if (!ctx)
+ return -ENOMEM;
+
+ *ctx = (struct bm_aio_ctx) {
.mdev = mdev,
.in_flight = ATOMIC_INIT(1),
- .done = COMPLETION_INITIALIZER_ONSTACK(ctx.done),
+ .done = 0,
.flags = BM_AIO_COPY_PAGES,
+ .error = 0,
+ .kref = { ATOMIC_INIT(2) },
};
- if (bm_test_page_unchanged(mdev->bitmap->bm_pages[idx])) {
- dynamic_dev_dbg(DEV, "skipped bm page write for idx %u\n", idx);
- return 0;
+ if (!get_ldev_if_state(mdev, D_ATTACHING)) { /* put is in bm_aio_ctx_destroy() */
+ dev_err(DEV, "ASSERT FAILED: get_ldev_if_state() == 1 in drbd_bm_write_page()\n");
+ kfree(ctx);
+ return -ENODEV;
}
- bm_page_io_async(&ctx, idx, WRITE_SYNC);
- wait_for_completion(&ctx.done);
+ bm_page_io_async(ctx, idx, WRITE_SYNC);
+ wait_until_done_or_disk_failure(mdev, mdev->ldev, &ctx->done);
- if (ctx.error)
+ if (ctx->error)
drbd_chk_io_error(mdev, 1, true);
/* that should force detach, so the in memory bitmap will be
* gone in a moment as well. */
mdev->bm_writ_cnt++;
- return ctx.error;
+ err = atomic_read(&ctx->in_flight) ? -EIO : ctx->error;
+ kref_put(&ctx->kref, &bm_aio_ctx_destroy);
+ return err;
}
/* NOTE
diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index 8d680562ba73..02f013a073a7 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -712,7 +712,6 @@ struct drbd_request {
struct list_head tl_requests; /* ring list in the transfer log */
struct bio *master_bio; /* master bio pointer */
unsigned long rq_state; /* see comments above _req_mod() */
- int seq_num;
unsigned long start_time;
};
@@ -851,6 +850,7 @@ enum {
NEW_CUR_UUID, /* Create new current UUID when thawing IO */
AL_SUSPENDED, /* Activity logging is currently suspended. */
AHEAD_TO_SYNC_SOURCE, /* Ahead -> SyncSource queued */
+ STATE_SENT, /* Do not change state/UUIDs while this is set */
};
struct drbd_bitmap; /* opaque for drbd_conf */
@@ -862,31 +862,30 @@ enum bm_flag {
BM_P_VMALLOCED = 0x10000, /* internal use only, will be masked out */
/* currently locked for bulk operation */
- BM_LOCKED_MASK = 0x7,
+ BM_LOCKED_MASK = 0xf,
/* in detail, that is: */
BM_DONT_CLEAR = 0x1,
BM_DONT_SET = 0x2,
BM_DONT_TEST = 0x4,
+ /* so we can mark it locked for bulk operation,
+ * and still allow all non-bulk operations */
+ BM_IS_LOCKED = 0x8,
+
/* (test bit, count bit) allowed (common case) */
- BM_LOCKED_TEST_ALLOWED = 0x3,
+ BM_LOCKED_TEST_ALLOWED = BM_DONT_CLEAR | BM_DONT_SET | BM_IS_LOCKED,
/* testing bits, as well as setting new bits allowed, but clearing bits
* would be unexpected. Used during bitmap receive. Setting new bits
* requires sending of "out-of-sync" information, though. */
- BM_LOCKED_SET_ALLOWED = 0x1,
+ BM_LOCKED_SET_ALLOWED = BM_DONT_CLEAR | BM_IS_LOCKED,
- /* clear is not expected while bitmap is locked for bulk operation */
+ /* for drbd_bm_write_copy_pages, everything is allowed,
+ * only concurrent bulk operations are locked out. */
+ BM_LOCKED_CHANGE_ALLOWED = BM_IS_LOCKED,
};
-
-/* TODO sort members for performance
- * MAYBE group them further */
-
-/* THINK maybe we actually want to use the default "event/%s" worker threads
- * or similar in linux 2.6, which uses per cpu data and threads.
- */
struct drbd_work_queue {
struct list_head q;
struct semaphore s; /* producers up it, worker down()s it */
@@ -938,8 +937,7 @@ struct drbd_backing_dev {
};
struct drbd_md_io {
- struct drbd_conf *mdev;
- struct completion event;
+ unsigned int done;
int error;
};
@@ -1022,6 +1020,7 @@ struct drbd_conf {
struct drbd_tl_epoch *newest_tle;
struct drbd_tl_epoch *oldest_tle;
struct list_head out_of_sequence_requests;
+ struct list_head barrier_acked_requests;
struct hlist_head *tl_hash;
unsigned int tl_hash_s;
@@ -1056,6 +1055,8 @@ struct drbd_conf {
struct crypto_hash *csums_tfm;
struct crypto_hash *verify_tfm;
+ unsigned long last_reattach_jif;
+ unsigned long last_reconnect_jif;
struct drbd_thread receiver;
struct drbd_thread worker;
struct drbd_thread asender;
@@ -1094,7 +1095,8 @@ struct drbd_conf {
wait_queue_head_t ee_wait;
struct page *md_io_page; /* one page buffer for md_io */
struct page *md_io_tmpp; /* for logical_block_size != 512 */
- struct mutex md_io_mutex; /* protects the md_io_buffer */
+ struct drbd_md_io md_io;
+ atomic_t md_io_in_use; /* protects the md_io, md_io_page and md_io_tmpp */
spinlock_t al_lock;
wait_queue_head_t al_wait;
struct lru_cache *act_log; /* activity log */
@@ -1228,8 +1230,8 @@ extern int drbd_send_uuids(struct drbd_conf *mdev);
extern int drbd_send_uuids_skip_initial_sync(struct drbd_conf *mdev);
extern int drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev);
extern int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags flags);
-extern int _drbd_send_state(struct drbd_conf *mdev);
-extern int drbd_send_state(struct drbd_conf *mdev);
+extern int drbd_send_state(struct drbd_conf *mdev, union drbd_state s);
+extern int drbd_send_current_state(struct drbd_conf *mdev);
extern int _drbd_send_cmd(struct drbd_conf *mdev, struct socket *sock,
enum drbd_packets cmd, struct p_header80 *h,
size_t size, unsigned msg_flags);
@@ -1461,6 +1463,7 @@ extern int drbd_bm_e_weight(struct drbd_conf *mdev, unsigned long enr);
extern int drbd_bm_write_page(struct drbd_conf *mdev, unsigned int idx) __must_hold(local);
extern int drbd_bm_read(struct drbd_conf *mdev) __must_hold(local);
extern int drbd_bm_write(struct drbd_conf *mdev) __must_hold(local);
+extern int drbd_bm_write_copy_pages(struct drbd_conf *mdev) __must_hold(local);
extern unsigned long drbd_bm_ALe_set_all(struct drbd_conf *mdev,
unsigned long al_enr);
extern size_t drbd_bm_words(struct drbd_conf *mdev);
@@ -1493,11 +1496,38 @@ extern struct kmem_cache *drbd_al_ext_cache; /* activity log extents */
extern mempool_t *drbd_request_mempool;
extern mempool_t *drbd_ee_mempool;
-extern struct page *drbd_pp_pool; /* drbd's page pool */
+/* drbd's page pool, used to buffer data received from the peer,
+ * or data requested by the peer.
+ *
+ * This does not have an emergency reserve.
+ *
+ * When allocating from this pool, it first takes pages from the pool.
+ * Only if the pool is depleted will try to allocate from the system.
+ *
+ * The assumption is that pages taken from this pool will be processed,
+ * and given back, "quickly", and then can be recycled, so we can avoid
+ * frequent calls to alloc_page(), and still will be able to make progress even
+ * under memory pressure.
+ */
+extern struct page *drbd_pp_pool;
extern spinlock_t drbd_pp_lock;
extern int drbd_pp_vacant;
extern wait_queue_head_t drbd_pp_wait;
+/* We also need a standard (emergency-reserve backed) page pool
+ * for meta data IO (activity log, bitmap).
+ * We can keep it global, as long as it is used as "N pages at a time".
+ * 128 should be plenty, currently we probably can get away with as few as 1.
+ */
+#define DRBD_MIN_POOL_PAGES 128
+extern mempool_t *drbd_md_io_page_pool;
+
+/* We also need to make sure we get a bio
+ * when we need it for housekeeping purposes */
+extern struct bio_set *drbd_md_io_bio_set;
+/* to allocate from that set */
+extern struct bio *bio_alloc_drbd(gfp_t gfp_mask);
+
extern rwlock_t global_state_lock;
extern struct drbd_conf *drbd_new_device(unsigned int minor);
@@ -1536,8 +1566,12 @@ extern void resume_next_sg(struct drbd_conf *mdev);
extern void suspend_other_sg(struct drbd_conf *mdev);
extern int drbd_resync_finished(struct drbd_conf *mdev);
/* maybe rather drbd_main.c ? */
+extern void *drbd_md_get_buffer(struct drbd_conf *mdev);
+extern void drbd_md_put_buffer(struct drbd_conf *mdev);
extern int drbd_md_sync_page_io(struct drbd_conf *mdev,
- struct drbd_backing_dev *bdev, sector_t sector, int rw);
+ struct drbd_backing_dev *bdev, sector_t sector, int rw);
+extern void wait_until_done_or_disk_failure(struct drbd_conf *mdev, struct drbd_backing_dev *bdev,
+ unsigned int *done);
extern void drbd_ov_oos_found(struct drbd_conf*, sector_t, int);
extern void drbd_rs_controller_reset(struct drbd_conf *mdev);
@@ -1754,19 +1788,6 @@ static inline struct page *page_chain_next(struct page *page)
#define page_chain_for_each_safe(page, n) \
for (; page && ({ n = page_chain_next(page); 1; }); page = n)
-static inline int drbd_bio_has_active_page(struct bio *bio)
-{
- struct bio_vec *bvec;
- int i;
-
- __bio_for_each_segment(bvec, bio, i, 0) {
- if (page_count(bvec->bv_page) > 1)
- return 1;
- }
-
- return 0;
-}
-
static inline int drbd_ee_has_active_page(struct drbd_epoch_entry *e)
{
struct page *page = e->pages;
@@ -1777,7 +1798,6 @@ static inline int drbd_ee_has_active_page(struct drbd_epoch_entry *e)
return 0;
}
-
static inline void drbd_state_lock(struct drbd_conf *mdev)
{
wait_event(mdev->misc_wait,
@@ -2230,7 +2250,7 @@ static inline void drbd_get_syncer_progress(struct drbd_conf *mdev,
* Note: currently we don't support such large bitmaps on 32bit
* arch anyways, but no harm done to be prepared for it here.
*/
- unsigned int shift = mdev->rs_total >= (1ULL << 32) ? 16 : 10;
+ unsigned int shift = mdev->rs_total > UINT_MAX ? 16 : 10;
unsigned long left = *bits_left >> shift;
unsigned long total = 1UL + (mdev->rs_total >> shift);
unsigned long tmp = 1000UL - left * 1000UL/total;
@@ -2306,12 +2326,12 @@ static inline int drbd_state_is_stable(struct drbd_conf *mdev)
case D_OUTDATED:
case D_CONSISTENT:
case D_UP_TO_DATE:
+ case D_FAILED:
/* disk state is stable as well. */
break;
/* no new io accepted during tansitional states */
case D_ATTACHING:
- case D_FAILED:
case D_NEGOTIATING:
case D_UNKNOWN:
case D_MASK:
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 211fc44f84be..920ede2829d6 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -139,6 +139,8 @@ struct kmem_cache *drbd_bm_ext_cache; /* bitmap extents */
struct kmem_cache *drbd_al_ext_cache; /* activity log extents */
mempool_t *drbd_request_mempool;
mempool_t *drbd_ee_mempool;
+mempool_t *drbd_md_io_page_pool;
+struct bio_set *drbd_md_io_bio_set;
/* I do not use a standard mempool, because:
1) I want to hand out the pre-allocated objects first.
@@ -159,7 +161,24 @@ static const struct block_device_operations drbd_ops = {
.release = drbd_release,
};
-#define ARRY_SIZE(A) (sizeof(A)/sizeof(A[0]))
+static void bio_destructor_drbd(struct bio *bio)
+{
+ bio_free(bio, drbd_md_io_bio_set);
+}
+
+struct bio *bio_alloc_drbd(gfp_t gfp_mask)
+{
+ struct bio *bio;
+
+ if (!drbd_md_io_bio_set)
+ return bio_alloc(gfp_mask, 1);
+
+ bio = bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set);
+ if (!bio)
+ return NULL;
+ bio->bi_destructor = bio_destructor_drbd;
+ return bio;
+}
#ifdef __CHECKER__
/* When checking with sparse, and this is an inline function, sparse will
@@ -208,6 +227,7 @@ static int tl_init(struct drbd_conf *mdev)
mdev->oldest_tle = b;
mdev->newest_tle = b;
INIT_LIST_HEAD(&mdev->out_of_sequence_requests);
+ INIT_LIST_HEAD(&mdev->barrier_acked_requests);
mdev->tl_hash = NULL;
mdev->tl_hash_s = 0;
@@ -246,9 +266,7 @@ void _tl_add_barrier(struct drbd_conf *mdev, struct drbd_tl_epoch *new)
new->n_writes = 0;
newest_before = mdev->newest_tle;
- /* never send a barrier number == 0, because that is special-cased
- * when using TCQ for our write ordering code */
- new->br_number = (newest_before->br_number+1) ?: 1;
+ new->br_number = newest_before->br_number+1;
if (mdev->newest_tle != new) {
mdev->newest_tle->next = new;
mdev->newest_tle = new;
@@ -311,7 +329,7 @@ void tl_release(struct drbd_conf *mdev, unsigned int barrier_nr,
These have been list_move'd to the out_of_sequence_requests list in
_req_mod(, barrier_acked) above.
*/
- list_del_init(&b->requests);
+ list_splice_init(&b->requests, &mdev->barrier_acked_requests);
nob = b->next;
if (test_and_clear_bit(CREATE_BARRIER, &mdev->flags)) {
@@ -411,6 +429,23 @@ static void _tl_restart(struct drbd_conf *mdev, enum drbd_req_event what)
b = tmp;
list_splice(&carry_reads, &b->requests);
}
+
+ /* Actions operating on the disk state, also want to work on
+ requests that got barrier acked. */
+ switch (what) {
+ case fail_frozen_disk_io:
+ case restart_frozen_disk_io:
+ list_for_each_safe(le, tle, &mdev->barrier_acked_requests) {
+ req = list_entry(le, struct drbd_request, tl_requests);
+ _req_mod(req, what);
+ }
+
+ case connection_lost_while_pending:
+ case resend:
+ break;
+ default:
+ dev_err(DEV, "what = %d in _tl_restart()\n", what);
+ }
}
@@ -458,6 +493,38 @@ void tl_restart(struct drbd_conf *mdev, enum drbd_req_event what)
}
/**
+ * tl_abort_disk_io() - Abort disk I/O for all requests for a certain mdev in the TL
+ * @mdev: DRBD device.
+ */
+void tl_abort_disk_io(struct drbd_conf *mdev)
+{
+ struct drbd_tl_epoch *b;
+ struct list_head *le, *tle;
+ struct drbd_request *req;
+
+ spin_lock_irq(&mdev->req_lock);
+ b = mdev->oldest_tle;
+ while (b) {
+ list_for_each_safe(le, tle, &b->requests) {
+ req = list_entry(le, struct drbd_request, tl_requests);
+ if (!(req->rq_state & RQ_LOCAL_PENDING))
+ continue;
+ _req_mod(req, abort_disk_io);
+ }
+ b = b->next;
+ }
+
+ list_for_each_safe(le, tle, &mdev->barrier_acked_requests) {
+ req = list_entry(le, struct drbd_request, tl_requests);
+ if (!(req->rq_state & RQ_LOCAL_PENDING))
+ continue;
+ _req_mod(req, abort_disk_io);
+ }
+
+ spin_unlock_irq(&mdev->req_lock);
+}
+
+/**
* cl_wide_st_chg() - true if the state change is a cluster wide one
* @mdev: DRBD device.
* @os: old (current) state.
@@ -470,7 +537,7 @@ static int cl_wide_st_chg(struct drbd_conf *mdev,
((os.role != R_PRIMARY && ns.role == R_PRIMARY) ||
(os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) ||
(os.conn != C_STARTING_SYNC_S && ns.conn == C_STARTING_SYNC_S) ||
- (os.disk != D_DISKLESS && ns.disk == D_DISKLESS))) ||
+ (os.disk != D_FAILED && ns.disk == D_FAILED))) ||
(os.conn >= C_CONNECTED && ns.conn == C_DISCONNECTING) ||
(os.conn == C_CONNECTED && ns.conn == C_VERIFY_S);
}
@@ -509,8 +576,16 @@ static enum drbd_state_rv is_valid_state(struct drbd_conf *, union drbd_state);
static enum drbd_state_rv is_valid_state_transition(struct drbd_conf *,
union drbd_state,
union drbd_state);
+enum sanitize_state_warnings {
+ NO_WARNING,
+ ABORTED_ONLINE_VERIFY,
+ ABORTED_RESYNC,
+ CONNECTION_LOST_NEGOTIATING,
+ IMPLICITLY_UPGRADED_DISK,
+ IMPLICITLY_UPGRADED_PDSK,
+};
static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state os,
- union drbd_state ns, const char **warn_sync_abort);
+ union drbd_state ns, enum sanitize_state_warnings *warn);
int drbd_send_state_req(struct drbd_conf *,
union drbd_state, union drbd_state);
@@ -785,6 +860,13 @@ is_valid_state_transition(struct drbd_conf *mdev, union drbd_state ns,
if (ns.conn == os.conn && ns.conn == C_WF_REPORT_PARAMS)
rv = SS_IN_TRANSIENT_STATE;
+ /* While establishing a connection only allow cstate to change.
+ Delay/refuse role changes, detach attach etc... */
+ if (test_bit(STATE_SENT, &mdev->flags) &&
+ !(os.conn == C_WF_REPORT_PARAMS ||
+ (ns.conn == C_WF_REPORT_PARAMS && os.conn == C_WF_CONNECTION)))
+ rv = SS_IN_TRANSIENT_STATE;
+
if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && os.conn < C_CONNECTED)
rv = SS_NEED_CONNECTION;
@@ -803,6 +885,21 @@ is_valid_state_transition(struct drbd_conf *mdev, union drbd_state ns,
return rv;
}
+static void print_sanitize_warnings(struct drbd_conf *mdev, enum sanitize_state_warnings warn)
+{
+ static const char *msg_table[] = {
+ [NO_WARNING] = "",
+ [ABORTED_ONLINE_VERIFY] = "Online-verify aborted.",
+ [ABORTED_RESYNC] = "Resync aborted.",
+ [CONNECTION_LOST_NEGOTIATING] = "Connection lost while negotiating, no data!",
+ [IMPLICITLY_UPGRADED_DISK] = "Implicitly upgraded disk",
+ [IMPLICITLY_UPGRADED_PDSK] = "Implicitly upgraded pdsk",
+ };
+
+ if (warn != NO_WARNING)
+ dev_warn(DEV, "%s\n", msg_table[warn]);
+}
+
/**
* sanitize_state() - Resolves implicitly necessary additional changes to a state transition
* @mdev: DRBD device.
@@ -814,11 +911,14 @@ is_valid_state_transition(struct drbd_conf *mdev, union drbd_state ns,
* to D_UNKNOWN. This rule and many more along those lines are in this function.
*/
static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state os,
- union drbd_state ns, const char **warn_sync_abort)
+ union drbd_state ns, enum sanitize_state_warnings *warn)
{
enum drbd_fencing_p fp;
enum drbd_disk_state disk_min, disk_max, pdsk_min, pdsk_max;
+ if (warn)
+ *warn = NO_WARNING;
+
fp = FP_DONT_CARE;
if (get_ldev(mdev)) {
fp = mdev->ldev->dc.fencing;
@@ -833,18 +933,13 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
/* After a network error (+C_TEAR_DOWN) only C_UNCONNECTED or C_DISCONNECTING can follow.
* If you try to go into some Sync* state, that shall fail (elsewhere). */
if (os.conn >= C_TIMEOUT && os.conn <= C_TEAR_DOWN &&
- ns.conn != C_UNCONNECTED && ns.conn != C_DISCONNECTING && ns.conn <= C_TEAR_DOWN)
+ ns.conn != C_UNCONNECTED && ns.conn != C_DISCONNECTING && ns.conn <= C_CONNECTED)
ns.conn = os.conn;
/* we cannot fail (again) if we already detached */
if (ns.disk == D_FAILED && os.disk == D_DISKLESS)
ns.disk = D_DISKLESS;
- /* if we are only D_ATTACHING yet,
- * we can (and should) go directly to D_DISKLESS. */
- if (ns.disk == D_FAILED && os.disk == D_ATTACHING)
- ns.disk = D_DISKLESS;
-
/* After C_DISCONNECTING only C_STANDALONE may follow */
if (os.conn == C_DISCONNECTING && ns.conn != C_STANDALONE)
ns.conn = os.conn;
@@ -863,10 +958,9 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
/* Abort resync if a disk fails/detaches */
if (os.conn > C_CONNECTED && ns.conn > C_CONNECTED &&
(ns.disk <= D_FAILED || ns.pdsk <= D_FAILED)) {
- if (warn_sync_abort)
- *warn_sync_abort =
- os.conn == C_VERIFY_S || os.conn == C_VERIFY_T ?
- "Online-verify" : "Resync";
+ if (warn)
+ *warn = os.conn == C_VERIFY_S || os.conn == C_VERIFY_T ?
+ ABORTED_ONLINE_VERIFY : ABORTED_RESYNC;
ns.conn = C_CONNECTED;
}
@@ -877,7 +971,8 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
ns.disk = mdev->new_state_tmp.disk;
ns.pdsk = mdev->new_state_tmp.pdsk;
} else {
- dev_alert(DEV, "Connection lost while negotiating, no data!\n");
+ if (warn)
+ *warn = CONNECTION_LOST_NEGOTIATING;
ns.disk = D_DISKLESS;
ns.pdsk = D_UNKNOWN;
}
@@ -959,16 +1054,16 @@ static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state
ns.disk = disk_max;
if (ns.disk < disk_min) {
- dev_warn(DEV, "Implicitly set disk from %s to %s\n",
- drbd_disk_str(ns.disk), drbd_disk_str(disk_min));
+ if (warn)
+ *warn = IMPLICITLY_UPGRADED_DISK;
ns.disk = disk_min;
}
if (ns.pdsk > pdsk_max)
ns.pdsk = pdsk_max;
if (ns.pdsk < pdsk_min) {
- dev_warn(DEV, "Implicitly set pdsk from %s to %s\n",
- drbd_disk_str(ns.pdsk), drbd_disk_str(pdsk_min));
+ if (warn)
+ *warn = IMPLICITLY_UPGRADED_PDSK;
ns.pdsk = pdsk_min;
}
@@ -1045,12 +1140,12 @@ __drbd_set_state(struct drbd_conf *mdev, union drbd_state ns,
{
union drbd_state os;
enum drbd_state_rv rv = SS_SUCCESS;
- const char *warn_sync_abort = NULL;
+ enum sanitize_state_warnings ssw;
struct after_state_chg_work *ascw;
os = mdev->state;
- ns = sanitize_state(mdev, os, ns, &warn_sync_abort);
+ ns = sanitize_state(mdev, os, ns, &ssw);
if (ns.i == os.i)
return SS_NOTHING_TO_DO;
@@ -1076,8 +1171,7 @@ __drbd_set_state(struct drbd_conf *mdev, union drbd_state ns,
return rv;
}
- if (warn_sync_abort)
- dev_warn(DEV, "%s aborted.\n", warn_sync_abort);
+ print_sanitize_warnings(mdev, ssw);
{
char *pbp, pb[300];
@@ -1243,7 +1337,7 @@ __drbd_set_state(struct drbd_conf *mdev, union drbd_state ns,
drbd_thread_stop_nowait(&mdev->receiver);
/* Upon network failure, we need to restart the receiver. */
- if (os.conn > C_TEAR_DOWN &&
+ if (os.conn > C_WF_CONNECTION &&
ns.conn <= C_TEAR_DOWN && ns.conn >= C_TIMEOUT)
drbd_thread_restart_nowait(&mdev->receiver);
@@ -1251,6 +1345,15 @@ __drbd_set_state(struct drbd_conf *mdev, union drbd_state ns,
if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED)
drbd_resume_al(mdev);
+ /* remember last connect and attach times so request_timer_fn() won't
+ * kill newly established sessions while we are still trying to thaw
+ * previously frozen IO */
+ if (os.conn != C_WF_REPORT_PARAMS && ns.conn == C_WF_REPORT_PARAMS)
+ mdev->last_reconnect_jif = jiffies;
+ if ((os.disk == D_ATTACHING || os.disk == D_NEGOTIATING) &&
+ ns.disk > D_NEGOTIATING)
+ mdev->last_reattach_jif = jiffies;
+
ascw = kmalloc(sizeof(*ascw), GFP_ATOMIC);
if (ascw) {
ascw->os = os;
@@ -1354,12 +1457,16 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
/* Here we have the actions that are performed after a
state change. This function might sleep */
+ if (os.disk <= D_NEGOTIATING && ns.disk > D_NEGOTIATING)
+ mod_timer(&mdev->request_timer, jiffies + HZ);
+
nsm.i = -1;
if (ns.susp_nod) {
if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED)
what = resend;
- if (os.disk == D_ATTACHING && ns.disk > D_ATTACHING)
+ if ((os.disk == D_ATTACHING || os.disk == D_NEGOTIATING) &&
+ ns.disk > D_NEGOTIATING)
what = restart_frozen_disk_io;
if (what != nothing)
@@ -1408,7 +1515,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
/* Do not change the order of the if above and the two below... */
if (os.pdsk == D_DISKLESS && ns.pdsk > D_DISKLESS) { /* attach on the peer */
drbd_send_uuids(mdev);
- drbd_send_state(mdev);
+ drbd_send_state(mdev, ns);
}
/* No point in queuing send_bitmap if we don't have a connection
* anymore, so check also the _current_ state, not only the new state
@@ -1441,11 +1548,11 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
}
if (ns.pdsk < D_INCONSISTENT && get_ldev(mdev)) {
- if (ns.peer == R_PRIMARY && mdev->ldev->md.uuid[UI_BITMAP] == 0) {
+ if (os.peer == R_SECONDARY && ns.peer == R_PRIMARY &&
+ mdev->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) {
drbd_uuid_new_current(mdev);
drbd_send_uuids(mdev);
}
-
/* D_DISKLESS Peer becomes secondary */
if (os.peer == R_PRIMARY && ns.peer == R_SECONDARY)
/* We may still be Primary ourselves.
@@ -1473,14 +1580,14 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
os.disk == D_ATTACHING && ns.disk == D_NEGOTIATING) {
drbd_send_sizes(mdev, 0, 0); /* to start sync... */
drbd_send_uuids(mdev);
- drbd_send_state(mdev);
+ drbd_send_state(mdev, ns);
}
/* We want to pause/continue resync, tell peer. */
if (ns.conn >= C_CONNECTED &&
((os.aftr_isp != ns.aftr_isp) ||
(os.user_isp != ns.user_isp)))
- drbd_send_state(mdev);
+ drbd_send_state(mdev, ns);
/* In case one of the isp bits got set, suspend other devices. */
if ((!os.aftr_isp && !os.peer_isp && !os.user_isp) &&
@@ -1490,10 +1597,10 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
/* Make sure the peer gets informed about eventual state
changes (ISP bits) while we were in WFReportParams. */
if (os.conn == C_WF_REPORT_PARAMS && ns.conn >= C_CONNECTED)
- drbd_send_state(mdev);
+ drbd_send_state(mdev, ns);
if (os.conn != C_AHEAD && ns.conn == C_AHEAD)
- drbd_send_state(mdev);
+ drbd_send_state(mdev, ns);
/* We are in the progress to start a full sync... */
if ((os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) ||
@@ -1513,33 +1620,38 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
/* first half of local IO error, failure to attach,
* or administrative detach */
if (os.disk != D_FAILED && ns.disk == D_FAILED) {
- enum drbd_io_error_p eh;
- int was_io_error;
+ enum drbd_io_error_p eh = EP_PASS_ON;
+ int was_io_error = 0;
/* corresponding get_ldev was in __drbd_set_state, to serialize
- * our cleanup here with the transition to D_DISKLESS,
- * so it is safe to dreference ldev here. */
- eh = mdev->ldev->dc.on_io_error;
- was_io_error = test_and_clear_bit(WAS_IO_ERROR, &mdev->flags);
-
- /* current state still has to be D_FAILED,
- * there is only one way out: to D_DISKLESS,
- * and that may only happen after our put_ldev below. */
- if (mdev->state.disk != D_FAILED)
- dev_err(DEV,
- "ASSERT FAILED: disk is %s during detach\n",
- drbd_disk_str(mdev->state.disk));
-
- if (drbd_send_state(mdev))
- dev_warn(DEV, "Notified peer that I am detaching my disk\n");
- else
- dev_err(DEV, "Sending state for detaching disk failed\n");
-
- drbd_rs_cancel_all(mdev);
-
- /* In case we want to get something to stable storage still,
- * this may be the last chance.
- * Following put_ldev may transition to D_DISKLESS. */
- drbd_md_sync(mdev);
+ * our cleanup here with the transition to D_DISKLESS.
+ * But is is still not save to dreference ldev here, since
+ * we might come from an failed Attach before ldev was set. */
+ if (mdev->ldev) {
+ 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);
+
+ /* current state still has to be D_FAILED,
+ * there is only one way out: to D_DISKLESS,
+ * and that may only happen after our put_ldev below. */
+ if (mdev->state.disk != D_FAILED)
+ dev_err(DEV,
+ "ASSERT FAILED: disk is %s during detach\n",
+ drbd_disk_str(mdev->state.disk));
+
+ if (ns.conn >= C_CONNECTED)
+ drbd_send_state(mdev, ns);
+
+ drbd_rs_cancel_all(mdev);
+
+ /* In case we want to get something to stable storage still,
+ * this may be the last chance.
+ * Following put_ldev may transition to D_DISKLESS. */
+ drbd_md_sync(mdev);
+ }
put_ldev(mdev);
if (was_io_error && eh == EP_CALL_HELPER)
@@ -1561,16 +1673,17 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
mdev->rs_failed = 0;
atomic_set(&mdev->rs_pending_cnt, 0);
- if (drbd_send_state(mdev))
- dev_warn(DEV, "Notified peer that I'm now diskless.\n");
+ if (ns.conn >= C_CONNECTED)
+ drbd_send_state(mdev, ns);
+
/* corresponding get_ldev in __drbd_set_state
* this may finally trigger drbd_ldev_destroy. */
put_ldev(mdev);
}
/* Notify peer that I had a local IO error, and did not detached.. */
- if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT)
- drbd_send_state(mdev);
+ if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT && ns.conn >= C_CONNECTED)
+ drbd_send_state(mdev, ns);
/* Disks got bigger while they were detached */
if (ns.disk > D_NEGOTIATING && ns.pdsk > D_NEGOTIATING &&
@@ -1588,7 +1701,13 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
/* sync target done with resync. Explicitly notify peer, even though
* it should (at least for non-empty resyncs) already know itself. */
if (os.disk < D_UP_TO_DATE && os.conn >= C_SYNC_SOURCE && ns.conn == C_CONNECTED)
- drbd_send_state(mdev);
+ drbd_send_state(mdev, ns);
+
+ /* Wake up role changes, that were delayed because of connection establishing */
+ if (os.conn == C_WF_REPORT_PARAMS && ns.conn != C_WF_REPORT_PARAMS) {
+ clear_bit(STATE_SENT, &mdev->flags);
+ wake_up(&mdev->state_wait);
+ }
/* This triggers bitmap writeout of potentially still unwritten pages
* if the resync finished cleanly, or aborted because of peer disk
@@ -1598,8 +1717,8 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
* No harm done if some bits change during this phase.
*/
if (os.conn > C_CONNECTED && ns.conn <= C_CONNECTED && get_ldev(mdev)) {
- drbd_queue_bitmap_io(mdev, &drbd_bm_write, NULL,
- "write from resync_finished", BM_LOCKED_SET_ALLOWED);
+ drbd_queue_bitmap_io(mdev, &drbd_bm_write_copy_pages, NULL,
+ "write from resync_finished", BM_LOCKED_CHANGE_ALLOWED);
put_ldev(mdev);
}
@@ -2057,7 +2176,11 @@ int drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev)
D_ASSERT(mdev->state.disk == D_UP_TO_DATE);
- uuid = mdev->ldev->md.uuid[UI_BITMAP] + UUID_NEW_BM_OFFSET;
+ uuid = mdev->ldev->md.uuid[UI_BITMAP];
+ if (uuid && uuid != UUID_JUST_CREATED)
+ uuid = uuid + UUID_NEW_BM_OFFSET;
+ else
+ get_random_bytes(&uuid, sizeof(u64));
drbd_uuid_set(mdev, UI_BITMAP, uuid);
drbd_print_uuids(mdev, "updated sync UUID");
drbd_md_sync(mdev);
@@ -2089,6 +2212,10 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl
max_bio_size = DRBD_MAX_BIO_SIZE; /* ... multiple BIOs per peer_request */
}
+ /* Never allow old drbd (up to 8.3.7) to see more than 32KiB */
+ if (mdev->agreed_pro_version <= 94)
+ max_bio_size = min_t(int, max_bio_size, DRBD_MAX_SIZE_H80_PACKET);
+
p.d_size = cpu_to_be64(d_size);
p.u_size = cpu_to_be64(u_size);
p.c_size = cpu_to_be64(trigger_reply ? 0 : drbd_get_capacity(mdev->this_bdev));
@@ -2102,10 +2229,10 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl
}
/**
- * drbd_send_state() - Sends the drbd state to the peer
+ * drbd_send_current_state() - Sends the drbd state to the peer
* @mdev: DRBD device.
*/
-int drbd_send_state(struct drbd_conf *mdev)
+int drbd_send_current_state(struct drbd_conf *mdev)
{
struct socket *sock;
struct p_state p;
@@ -2131,6 +2258,37 @@ int drbd_send_state(struct drbd_conf *mdev)
return ok;
}
+/**
+ * drbd_send_state() - After a state change, sends the new state to the peer
+ * @mdev: DRBD device.
+ * @state: the state to send, not necessarily the current state.
+ *
+ * Each state change queues an "after_state_ch" work, which will eventually
+ * send the resulting new state to the peer. If more state changes happen
+ * between queuing and processing of the after_state_ch work, we still
+ * want to send each intermediary state in the order it occurred.
+ */
+int drbd_send_state(struct drbd_conf *mdev, union drbd_state state)
+{
+ struct socket *sock;
+ struct p_state p;
+ int ok = 0;
+
+ mutex_lock(&mdev->data.mutex);
+
+ p.state = cpu_to_be32(state.i);
+ sock = mdev->data.socket;
+
+ if (likely(sock != NULL)) {
+ ok = _drbd_send_cmd(mdev, sock, P_STATE,
+ (struct p_header80 *)&p, sizeof(p), 0);
+ }
+
+ mutex_unlock(&mdev->data.mutex);
+
+ return ok;
+}
+
int drbd_send_state_req(struct drbd_conf *mdev,
union drbd_state mask, union drbd_state val)
{
@@ -2615,7 +2773,7 @@ static int _drbd_send_bio(struct drbd_conf *mdev, struct bio *bio)
struct bio_vec *bvec;
int i;
/* hint all but last page with MSG_MORE */
- __bio_for_each_segment(bvec, bio, i, 0) {
+ bio_for_each_segment(bvec, bio, i) {
if (!_drbd_no_send_page(mdev, bvec->bv_page,
bvec->bv_offset, bvec->bv_len,
i == bio->bi_vcnt -1 ? 0 : MSG_MORE))
@@ -2629,7 +2787,7 @@ static int _drbd_send_zc_bio(struct drbd_conf *mdev, struct bio *bio)
struct bio_vec *bvec;
int i;
/* hint all but last page with MSG_MORE */
- __bio_for_each_segment(bvec, bio, i, 0) {
+ bio_for_each_segment(bvec, bio, i) {
if (!_drbd_send_page(mdev, bvec->bv_page,
bvec->bv_offset, bvec->bv_len,
i == bio->bi_vcnt -1 ? 0 : MSG_MORE))
@@ -2695,8 +2853,7 @@ int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req)
p.sector = cpu_to_be64(req->sector);
p.block_id = (unsigned long)req;
- p.seq_num = cpu_to_be32(req->seq_num =
- atomic_add_return(1, &mdev->packet_seq));
+ p.seq_num = cpu_to_be32(atomic_add_return(1, &mdev->packet_seq));
dp_flags = bio_flags_to_wire(mdev, req->master_bio->bi_rw);
@@ -2987,8 +3144,8 @@ void drbd_init_set_defaults(struct drbd_conf *mdev)
atomic_set(&mdev->rs_sect_in, 0);
atomic_set(&mdev->rs_sect_ev, 0);
atomic_set(&mdev->ap_in_flight, 0);
+ atomic_set(&mdev->md_io_in_use, 0);
- mutex_init(&mdev->md_io_mutex);
mutex_init(&mdev->data.mutex);
mutex_init(&mdev->meta.mutex);
sema_init(&mdev->data.work.s, 0);
@@ -3126,6 +3283,10 @@ static void drbd_destroy_mempools(void)
/* D_ASSERT(atomic_read(&drbd_pp_vacant)==0); */
+ if (drbd_md_io_bio_set)
+ bioset_free(drbd_md_io_bio_set);
+ if (drbd_md_io_page_pool)
+ mempool_destroy(drbd_md_io_page_pool);
if (drbd_ee_mempool)
mempool_destroy(drbd_ee_mempool);
if (drbd_request_mempool)
@@ -3139,6 +3300,8 @@ static void drbd_destroy_mempools(void)
if (drbd_al_ext_cache)
kmem_cache_destroy(drbd_al_ext_cache);
+ drbd_md_io_bio_set = NULL;
+ drbd_md_io_page_pool = NULL;
drbd_ee_mempool = NULL;
drbd_request_mempool = NULL;
drbd_ee_cache = NULL;
@@ -3162,6 +3325,8 @@ static int drbd_create_mempools(void)
drbd_bm_ext_cache = NULL;
drbd_al_ext_cache = NULL;
drbd_pp_pool = NULL;
+ drbd_md_io_page_pool = NULL;
+ drbd_md_io_bio_set = NULL;
/* caches */
drbd_request_cache = kmem_cache_create(
@@ -3185,6 +3350,16 @@ static int drbd_create_mempools(void)
goto Enomem;
/* mempools */
+#ifdef COMPAT_HAVE_BIOSET_CREATE
+ drbd_md_io_bio_set = bioset_create(DRBD_MIN_POOL_PAGES, 0);
+ if (drbd_md_io_bio_set == NULL)
+ goto Enomem;
+#endif
+
+ drbd_md_io_page_pool = mempool_create_page_pool(DRBD_MIN_POOL_PAGES, 0);
+ if (drbd_md_io_page_pool == NULL)
+ goto Enomem;
+
drbd_request_mempool = mempool_create(number,
mempool_alloc_slab, mempool_free_slab, drbd_request_cache);
if (drbd_request_mempool == NULL)
@@ -3262,6 +3437,8 @@ static void drbd_delete_device(unsigned int minor)
if (!mdev)
return;
+ del_timer_sync(&mdev->request_timer);
+
/* paranoia asserts */
if (mdev->open_cnt != 0)
dev_err(DEV, "open_cnt = %d in %s:%u", mdev->open_cnt,
@@ -3666,8 +3843,10 @@ void drbd_md_sync(struct drbd_conf *mdev)
if (!get_ldev_if_state(mdev, D_FAILED))
return;
- mutex_lock(&mdev->md_io_mutex);
- buffer = (struct meta_data_on_disk *)page_address(mdev->md_io_page);
+ buffer = drbd_md_get_buffer(mdev);
+ if (!buffer)
+ goto out;
+
memset(buffer, 0, 512);
buffer->la_size = cpu_to_be64(drbd_get_capacity(mdev->this_bdev));
@@ -3698,7 +3877,8 @@ void drbd_md_sync(struct drbd_conf *mdev)
* since we updated it on metadata. */
mdev->ldev->md.la_size_sect = drbd_get_capacity(mdev->this_bdev);
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
+out:
put_ldev(mdev);
}
@@ -3718,8 +3898,9 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
if (!get_ldev_if_state(mdev, D_ATTACHING))
return ERR_IO_MD_DISK;
- mutex_lock(&mdev->md_io_mutex);
- buffer = (struct meta_data_on_disk *)page_address(mdev->md_io_page);
+ buffer = drbd_md_get_buffer(mdev);
+ if (!buffer)
+ goto out;
if (!drbd_md_sync_page_io(mdev, bdev, bdev->md.md_offset, READ)) {
/* NOTE: can't do normal error processing here as this is
@@ -3780,7 +3961,8 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev)
mdev->sync_conf.al_extents = 127;
err:
- mutex_unlock(&mdev->md_io_mutex);
+ drbd_md_put_buffer(mdev);
+ out:
put_ldev(mdev);
return rv;
@@ -4183,12 +4365,11 @@ const char *drbd_buildtag(void)
static char buildtag[38] = "\0uilt-in";
if (buildtag[0] == 0) {
-#ifdef CONFIG_MODULES
- if (THIS_MODULE != NULL)
- sprintf(buildtag, "srcversion: %-24s", THIS_MODULE->srcversion);
- else
+#ifdef MODULE
+ sprintf(buildtag, "srcversion: %-24s", THIS_MODULE->srcversion);
+#else
+ buildtag[0] = 'b';
#endif
- buildtag[0] = 'b';
}
return buildtag;
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 946166e13953..6d4de6a72e80 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -289,7 +289,7 @@ static int _try_outdate_peer_async(void *data)
*/
spin_lock_irq(&mdev->req_lock);
ns = mdev->state;
- if (ns.conn < C_WF_REPORT_PARAMS) {
+ if (ns.conn < C_WF_REPORT_PARAMS && !test_bit(STATE_SENT, &mdev->flags)) {
ns.pdsk = nps;
_drbd_set_state(mdev, ns, CS_VERBOSE, NULL);
}
@@ -432,7 +432,7 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force)
/* if this was forced, we should consider sync */
if (forced)
drbd_send_uuids(mdev);
- drbd_send_state(mdev);
+ drbd_send_current_state(mdev);
}
drbd_md_sync(mdev);
@@ -845,9 +845,10 @@ void drbd_reconsider_max_bio_size(struct drbd_conf *mdev)
Because new from 8.3.8 onwards the peer can use multiple
BIOs for a single peer_request */
if (mdev->state.conn >= C_CONNECTED) {
- if (mdev->agreed_pro_version < 94)
- peer = mdev->peer_max_bio_size;
- else if (mdev->agreed_pro_version == 94)
+ if (mdev->agreed_pro_version < 94) {
+ peer = min_t(int, mdev->peer_max_bio_size, DRBD_MAX_SIZE_H80_PACKET);
+ /* Correct old drbd (up to 8.3.7) if it believes it can do more than 32KiB */
+ } else if (mdev->agreed_pro_version == 94)
peer = DRBD_MAX_SIZE_H80_PACKET;
else /* drbd 8.3.8 onwards */
peer = DRBD_MAX_BIO_SIZE;
@@ -1032,7 +1033,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
dev_err(DEV, "max capacity %llu smaller than disk size %llu\n",
(unsigned long long) drbd_get_max_capacity(nbc),
(unsigned long long) nbc->dc.disk_size);
- retcode = ERR_DISK_TO_SMALL;
+ retcode = ERR_DISK_TOO_SMALL;
goto fail;
}
@@ -1046,7 +1047,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
}
if (drbd_get_capacity(nbc->md_bdev) < min_md_device_sectors) {
- retcode = ERR_MD_DISK_TO_SMALL;
+ retcode = ERR_MD_DISK_TOO_SMALL;
dev_warn(DEV, "refusing attach: md-device too small, "
"at least %llu sectors needed for this meta-disk type\n",
(unsigned long long) min_md_device_sectors);
@@ -1057,7 +1058,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
* (we may currently be R_PRIMARY with no local disk...) */
if (drbd_get_max_capacity(nbc) <
drbd_get_capacity(mdev->this_bdev)) {
- retcode = ERR_DISK_TO_SMALL;
+ retcode = ERR_DISK_TOO_SMALL;
goto fail;
}
@@ -1138,7 +1139,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
if (drbd_md_test_flag(nbc, MDF_CONSISTENT) &&
drbd_new_dev_size(mdev, nbc, 0) < nbc->md.la_size_sect) {
dev_warn(DEV, "refusing to truncate a consistent device\n");
- retcode = ERR_DISK_TO_SMALL;
+ retcode = ERR_DISK_TOO_SMALL;
goto force_diskless_dec;
}
@@ -1336,17 +1337,34 @@ static int drbd_nl_detach(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
{
enum drbd_ret_code retcode;
int ret;
+ struct detach dt = {};
+
+ if (!detach_from_tags(mdev, nlp->tag_list, &dt)) {
+ reply->ret_code = ERR_MANDATORY_TAG;
+ goto out;
+ }
+
+ if (dt.detach_force) {
+ drbd_force_state(mdev, NS(disk, D_FAILED));
+ reply->ret_code = SS_SUCCESS;
+ goto out;
+ }
+
drbd_suspend_io(mdev); /* so no-one is stuck in drbd_al_begin_io */
+ drbd_md_get_buffer(mdev); /* make sure there is no in-flight meta-data IO */
retcode = drbd_request_state(mdev, NS(disk, D_FAILED));
+ drbd_md_put_buffer(mdev);
/* D_FAILED will transition to DISKLESS. */
ret = wait_event_interruptible(mdev->misc_wait,
mdev->state.disk != D_FAILED);
drbd_resume_io(mdev);
+
if ((int)retcode == (int)SS_IS_DISKLESS)
retcode = SS_NOTHING_TO_DO;
if (ret)
retcode = ERR_INTR;
reply->ret_code = retcode;
+out:
return 0;
}
@@ -1711,7 +1729,7 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
if (rs.no_resync && mdev->agreed_pro_version < 93) {
retcode = ERR_NEED_APV_93;
- goto fail;
+ goto fail_ldev;
}
if (mdev->ldev->known_size != drbd_get_capacity(mdev->ldev->backing_bdev))
@@ -1738,6 +1756,10 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
fail:
reply->ret_code = retcode;
return 0;
+
+ fail_ldev:
+ put_ldev(mdev);
+ goto fail;
}
static int drbd_nl_syncer_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
@@ -1941,6 +1963,7 @@ static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl
/* If there is still bitmap IO pending, probably because of a previous
* resync just being finished, wait for it before requesting a new resync. */
+ drbd_suspend_io(mdev);
wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
retcode = _drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T), CS_ORDERED);
@@ -1959,6 +1982,7 @@ static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl
retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T));
}
+ drbd_resume_io(mdev);
reply->ret_code = retcode;
return 0;
@@ -1980,6 +2004,7 @@ static int drbd_nl_invalidate_peer(struct drbd_conf *mdev, struct drbd_nl_cfg_re
/* If there is still bitmap IO pending, probably because of a previous
* resync just being finished, wait for it before requesting a new resync. */
+ drbd_suspend_io(mdev);
wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
retcode = _drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S), CS_ORDERED);
@@ -1998,6 +2023,7 @@ static int drbd_nl_invalidate_peer(struct drbd_conf *mdev, struct drbd_nl_cfg_re
} else
retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S));
}
+ drbd_resume_io(mdev);
reply->ret_code = retcode;
return 0;
@@ -2170,11 +2196,13 @@ static int drbd_nl_start_ov(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
/* If there is still bitmap IO pending, e.g. previous resync or verify
* just being finished, wait for it before requesting a new resync. */
+ drbd_suspend_io(mdev);
wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
/* w_make_ov_request expects position to be aligned */
mdev->ov_start_sector = args.start_sector & ~BM_SECT_PER_BIT;
reply->ret_code = drbd_request_state(mdev,NS(conn,C_VERIFY_S));
+ drbd_resume_io(mdev);
return 0;
}
diff --git a/drivers/block/drbd/drbd_proc.c b/drivers/block/drbd/drbd_proc.c
index 2959cdfb77f5..869bada2ed06 100644
--- a/drivers/block/drbd/drbd_proc.c
+++ b/drivers/block/drbd/drbd_proc.c
@@ -52,7 +52,7 @@ void seq_printf_with_thousands_grouping(struct seq_file *seq, long v)
if (unlikely(v >= 1000000)) {
/* cool: > GiByte/s */
seq_printf(seq, "%ld,", v / 1000000);
- v /= 1000000;
+ v %= 1000000;
seq_printf(seq, "%03ld,%03ld", v/1000, v % 1000);
} else if (likely(v >= 1000))
seq_printf(seq, "%ld,%03ld", v/1000, v % 1000);
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 43beaca53179..ea4836e0ae98 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -466,6 +466,7 @@ static int drbd_accept(struct drbd_conf *mdev, const char **what,
goto out;
}
(*newsock)->ops = sock->ops;
+ __module_get((*newsock)->ops->owner);
out:
return err;
@@ -664,7 +665,7 @@ static struct socket *drbd_wait_for_connect(struct drbd_conf *mdev)
timeo = mdev->net_conf->try_connect_int * HZ;
timeo += (random32() & 1) ? timeo / 7 : -timeo / 7; /* 28.5% random jitter */
- s_listen->sk->sk_reuse = 1; /* SO_REUSEADDR */
+ s_listen->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */
s_listen->sk->sk_rcvtimeo = timeo;
s_listen->sk->sk_sndtimeo = timeo;
drbd_setbufsize(s_listen, mdev->net_conf->sndbuf_size,
@@ -750,6 +751,7 @@ static int drbd_connect(struct drbd_conf *mdev)
{
struct socket *s, *sock, *msock;
int try, h, ok;
+ enum drbd_state_rv rv;
D_ASSERT(!mdev->data.socket);
@@ -841,8 +843,8 @@ retry:
}
} while (1);
- msock->sk->sk_reuse = 1; /* SO_REUSEADDR */
- sock->sk->sk_reuse = 1; /* SO_REUSEADDR */
+ msock->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */
+ sock->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */
sock->sk->sk_allocation = GFP_NOIO;
msock->sk->sk_allocation = GFP_NOIO;
@@ -888,25 +890,32 @@ retry:
}
}
- if (drbd_request_state(mdev, NS(conn, C_WF_REPORT_PARAMS)) < SS_SUCCESS)
- return 0;
-
sock->sk->sk_sndtimeo = mdev->net_conf->timeout*HZ/10;
sock->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT;
atomic_set(&mdev->packet_seq, 0);
mdev->peer_seq = 0;
- drbd_thread_start(&mdev->asender);
-
if (drbd_send_protocol(mdev) == -1)
return -1;
+ set_bit(STATE_SENT, &mdev->flags);
drbd_send_sync_param(mdev, &mdev->sync_conf);
drbd_send_sizes(mdev, 0, 0);
drbd_send_uuids(mdev);
- drbd_send_state(mdev);
+ drbd_send_current_state(mdev);
clear_bit(USE_DEGR_WFC_T, &mdev->flags);
clear_bit(RESIZE_PENDING, &mdev->flags);
+
+ spin_lock_irq(&mdev->req_lock);
+ rv = _drbd_set_state(_NS(mdev, conn, C_WF_REPORT_PARAMS), CS_VERBOSE, NULL);
+ if (mdev->state.conn != C_WF_REPORT_PARAMS)
+ clear_bit(STATE_SENT, &mdev->flags);
+ spin_unlock_irq(&mdev->req_lock);
+
+ if (rv < SS_SUCCESS)
+ return 0;
+
+ drbd_thread_start(&mdev->asender);
mod_timer(&mdev->request_timer, jiffies + HZ); /* just start it here. */
return 1;
@@ -957,7 +966,7 @@ static void drbd_flush(struct drbd_conf *mdev)
rv = blkdev_issue_flush(mdev->ldev->backing_bdev, GFP_KERNEL,
NULL);
if (rv) {
- dev_err(DEV, "local disk flush failed with status %d\n", rv);
+ dev_info(DEV, "local disk flush failed with status %d\n", rv);
/* would rather check on EOPNOTSUPP, but that is not reliable.
* don't try again for ANY return value != 0
* if (rv == -EOPNOTSUPP) */
@@ -1001,13 +1010,14 @@ static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev,
if (epoch_size != 0 &&
atomic_read(&epoch->active) == 0 &&
- test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags)) {
+ (test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags) || ev & EV_CLEANUP)) {
if (!(ev & EV_CLEANUP)) {
spin_unlock(&mdev->epoch_lock);
drbd_send_b_ack(mdev, epoch->barrier_nr, epoch_size);
spin_lock(&mdev->epoch_lock);
}
- dec_unacked(mdev);
+ if (test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags))
+ dec_unacked(mdev);
if (mdev->current_epoch != epoch) {
next_epoch = list_entry(epoch->list.next, struct drbd_epoch, list);
@@ -1096,7 +1106,11 @@ int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
/* In most cases, we will only need one bio. But in case the lower
* level restrictions happen to be different at this offset on this
* side than those of the sending peer, we may need to submit the
- * request in more than one bio. */
+ * request in more than one bio.
+ *
+ * Plain bio_alloc is good enough here, this is no DRBD internally
+ * generated bio, but a bio allocated on behalf of the peer.
+ */
next_bio:
bio = bio_alloc(GFP_NOIO, nr_pages);
if (!bio) {
@@ -1583,6 +1597,24 @@ static int e_send_discard_ack(struct drbd_conf *mdev, struct drbd_work *w, int u
return ok;
}
+static bool overlapping_resync_write(struct drbd_conf *mdev, struct drbd_epoch_entry *data_e)
+{
+
+ struct drbd_epoch_entry *rs_e;
+ bool rv = 0;
+
+ spin_lock_irq(&mdev->req_lock);
+ list_for_each_entry(rs_e, &mdev->sync_ee, w.list) {
+ if (overlaps(data_e->sector, data_e->size, rs_e->sector, rs_e->size)) {
+ rv = 1;
+ break;
+ }
+ }
+ spin_unlock_irq(&mdev->req_lock);
+
+ return rv;
+}
+
/* Called from receive_Data.
* Synchronize packets on sock with packets on msock.
*
@@ -1826,6 +1858,9 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
list_add(&e->w.list, &mdev->active_ee);
spin_unlock_irq(&mdev->req_lock);
+ if (mdev->state.conn == C_SYNC_TARGET)
+ wait_event(mdev->ee_wait, !overlapping_resync_write(mdev, e));
+
switch (mdev->net_conf->wire_protocol) {
case DRBD_PROT_C:
inc_unacked(mdev);
@@ -2420,7 +2455,7 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l
mdev->p_uuid[UI_BITMAP] = mdev->p_uuid[UI_HISTORY_START];
mdev->p_uuid[UI_HISTORY_START] = mdev->p_uuid[UI_HISTORY_START + 1];
- dev_info(DEV, "Did not got last syncUUID packet, corrected:\n");
+ dev_info(DEV, "Lost last syncUUID packet, corrected:\n");
drbd_uuid_dump(mdev, "peer", mdev->p_uuid, mdev->p_uuid[UI_SIZE], mdev->p_uuid[UI_FLAGS]);
return -1;
@@ -2806,10 +2841,10 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi
if (apv >= 88) {
if (apv == 88) {
- if (data_size > SHARED_SECRET_MAX) {
- dev_err(DEV, "verify-alg too long, "
- "peer wants %u, accepting only %u byte\n",
- data_size, SHARED_SECRET_MAX);
+ if (data_size > SHARED_SECRET_MAX || data_size == 0) {
+ dev_err(DEV, "verify-alg of wrong size, "
+ "peer wants %u, accepting only up to %u byte\n",
+ data_size, SHARED_SECRET_MAX);
return false;
}
@@ -3168,9 +3203,20 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
os = ns = mdev->state;
spin_unlock_irq(&mdev->req_lock);
- /* peer says his disk is uptodate, while we think it is inconsistent,
- * and this happens while we think we have a sync going on. */
- if (os.pdsk == D_INCONSISTENT && real_peer_disk == D_UP_TO_DATE &&
+ /* If some other part of the code (asender thread, timeout)
+ * already decided to close the connection again,
+ * we must not "re-establish" it here. */
+ if (os.conn <= C_TEAR_DOWN)
+ return false;
+
+ /* If this is the "end of sync" confirmation, usually the peer disk
+ * transitions from D_INCONSISTENT to D_UP_TO_DATE. For empty (0 bits
+ * set) resync started in PausedSyncT, or if the timing of pause-/
+ * unpause-sync events has been "just right", the peer disk may
+ * transition from D_CONSISTENT to D_UP_TO_DATE as well.
+ */
+ if ((os.pdsk == D_INCONSISTENT || os.pdsk == D_CONSISTENT) &&
+ real_peer_disk == D_UP_TO_DATE &&
os.conn > C_CONNECTED && os.disk == D_UP_TO_DATE) {
/* If we are (becoming) SyncSource, but peer is still in sync
* preparation, ignore its uptodate-ness to avoid flapping, it
@@ -3288,7 +3334,7 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
/* Nowadays only used when forcing a node into primary role and
setting its disk to UpToDate with that */
drbd_send_uuids(mdev);
- drbd_send_state(mdev);
+ drbd_send_current_state(mdev);
}
}
@@ -3776,6 +3822,13 @@ static void drbd_disconnect(struct drbd_conf *mdev)
if (mdev->state.conn == C_STANDALONE)
return;
+ /* We are about to start the cleanup after connection loss.
+ * Make sure drbd_make_request knows about that.
+ * Usually we should be in some network failure state already,
+ * but just in case we are not, we fix it up here.
+ */
+ drbd_force_state(mdev, NS(conn, C_NETWORK_FAILURE));
+
/* asender does not clean up anything. it must not interfere, either */
drbd_thread_stop(&mdev->asender);
drbd_free_sock(mdev);
@@ -3803,8 +3856,6 @@ static void drbd_disconnect(struct drbd_conf *mdev)
atomic_set(&mdev->rs_pending_cnt, 0);
wake_up(&mdev->misc_wait);
- del_timer(&mdev->request_timer);
-
/* make sure syncer is stopped and w_resume_next_sg queued */
del_timer_sync(&mdev->resync_timer);
resync_timer_fn((unsigned long)mdev);
@@ -4433,7 +4484,7 @@ static int got_BarrierAck(struct drbd_conf *mdev, struct p_header80 *h)
if (mdev->state.conn == C_AHEAD &&
atomic_read(&mdev->ap_in_flight) == 0 &&
- !test_and_set_bit(AHEAD_TO_SYNC_SOURCE, &mdev->current_epoch->flags)) {
+ !test_and_set_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags)) {
mdev->start_resync_timer.expires = jiffies + HZ;
add_timer(&mdev->start_resync_timer);
}
diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c
index 4a0f314086e5..9c5c84946b05 100644
--- a/drivers/block/drbd/drbd_req.c
+++ b/drivers/block/drbd/drbd_req.c
@@ -37,6 +37,7 @@ static void _drbd_start_io_acct(struct drbd_conf *mdev, struct drbd_request *req
const int rw = bio_data_dir(bio);
int cpu;
cpu = part_stat_lock();
+ part_round_stats(cpu, &mdev->vdisk->part0);
part_stat_inc(cpu, &mdev->vdisk->part0, ios[rw]);
part_stat_add(cpu, &mdev->vdisk->part0, sectors[rw], bio_sectors(bio));
part_inc_in_flight(&mdev->vdisk->part0, rw);
@@ -214,8 +215,7 @@ void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m)
{
const unsigned long s = req->rq_state;
struct drbd_conf *mdev = req->mdev;
- /* only WRITES may end up here without a master bio (on barrier ack) */
- int rw = req->master_bio ? bio_data_dir(req->master_bio) : WRITE;
+ int rw = req->rq_state & RQ_WRITE ? WRITE : READ;
/* we must not complete the master bio, while it is
* still being processed by _drbd_send_zc_bio (drbd_send_dblock)
@@ -230,7 +230,7 @@ void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m)
return;
if (s & RQ_NET_PENDING)
return;
- if (s & RQ_LOCAL_PENDING)
+ if (s & RQ_LOCAL_PENDING && !(s & RQ_LOCAL_ABORTED))
return;
if (req->master_bio) {
@@ -277,6 +277,9 @@ void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m)
req->master_bio = NULL;
}
+ if (s & RQ_LOCAL_PENDING)
+ return;
+
if ((s & RQ_NET_MASK) == 0 || (s & RQ_NET_DONE)) {
/* this is disconnected (local only) operation,
* or protocol C P_WRITE_ACK,
@@ -429,7 +432,7 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
break;
case completed_ok:
- if (bio_data_dir(req->master_bio) == WRITE)
+ if (req->rq_state & RQ_WRITE)
mdev->writ_cnt += req->size>>9;
else
mdev->read_cnt += req->size>>9;
@@ -438,7 +441,14 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
req->rq_state &= ~RQ_LOCAL_PENDING;
_req_may_be_done_not_susp(req, m);
- put_ldev(mdev);
+ break;
+
+ case abort_disk_io:
+ req->rq_state |= RQ_LOCAL_ABORTED;
+ if (req->rq_state & RQ_WRITE)
+ _req_may_be_done_not_susp(req, m);
+ else
+ goto goto_queue_for_net_read;
break;
case write_completed_with_error:
@@ -447,7 +457,6 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
__drbd_chk_io_error(mdev, false);
_req_may_be_done_not_susp(req, m);
- put_ldev(mdev);
break;
case read_ahead_completed_with_error:
@@ -455,7 +464,6 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
req->rq_state |= RQ_LOCAL_COMPLETED;
req->rq_state &= ~RQ_LOCAL_PENDING;
_req_may_be_done_not_susp(req, m);
- put_ldev(mdev);
break;
case read_completed_with_error:
@@ -467,7 +475,8 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
D_ASSERT(!(req->rq_state & RQ_NET_MASK));
__drbd_chk_io_error(mdev, false);
- put_ldev(mdev);
+
+ goto_queue_for_net_read:
/* no point in retrying if there is no good remote data,
* or we have no connection. */
@@ -556,10 +565,8 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
drbd_queue_work(&mdev->data.work, &req->w);
break;
- case oos_handed_to_network:
- /* actually the same */
+ case read_retry_remote_canceled:
case send_canceled:
- /* treat it the same */
case send_failed:
/* real cleanup will be done from tl_clear. just update flags
* so it is no longer marked as on the worker queue */
@@ -589,17 +596,17 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
}
req->rq_state &= ~RQ_NET_QUEUED;
req->rq_state |= RQ_NET_SENT;
- /* because _drbd_send_zc_bio could sleep, and may want to
- * dereference the bio even after the "write_acked_by_peer" and
- * "completed_ok" events came in, once we return from
- * _drbd_send_zc_bio (drbd_send_dblock), we have to check
- * whether it is done already, and end it. */
_req_may_be_done_not_susp(req, m);
break;
- case read_retry_remote_canceled:
+ case oos_handed_to_network:
+ /* Was not set PENDING, no longer QUEUED, so is now DONE
+ * as far as this connection is concerned. */
req->rq_state &= ~RQ_NET_QUEUED;
- /* fall through, in case we raced with drbd_disconnect */
+ req->rq_state |= RQ_NET_DONE;
+ _req_may_be_done_not_susp(req, m);
+ break;
+
case connection_lost_while_pending:
/* transfer log cleanup after connection loss */
/* assert something? */
@@ -616,8 +623,6 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
_req_may_be_done(req, m); /* Allowed while state.susp */
break;
- case write_acked_by_peer_and_sis:
- req->rq_state |= RQ_NET_SIS;
case conflict_discarded_by_peer:
/* for discarded conflicting writes of multiple primaries,
* there is no need to keep anything in the tl, potential
@@ -628,18 +633,15 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what,
(unsigned long long)req->sector, req->size);
req->rq_state |= RQ_NET_DONE;
/* fall through */
+ case write_acked_by_peer_and_sis:
case write_acked_by_peer:
+ if (what == write_acked_by_peer_and_sis)
+ req->rq_state |= RQ_NET_SIS;
/* protocol C; successfully written on peer.
- * Nothing to do here.
+ * Nothing more to do here.
* We want to keep the tl in place for all protocols, to cater
- * for volatile write-back caches on lower level devices.
- *
- * A barrier request is expected to have forced all prior
- * requests onto stable storage, so completion of a barrier
- * request could set NET_DONE right here, and not wait for the
- * P_BARRIER_ACK, but that is an unnecessary optimization. */
+ * for volatile write-back caches on lower level devices. */
- /* this makes it effectively the same as for: */
case recv_acked_by_peer:
/* protocol B; pretends to be successfully written on peer.
* see also notes above in handed_over_to_network about
@@ -773,6 +775,7 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, uns
int local, remote, send_oos = 0;
int err = -EIO;
int ret = 0;
+ union drbd_state s;
/* allocate outside of all locks; */
req = drbd_req_new(mdev, bio);
@@ -834,8 +837,9 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, uns
drbd_al_begin_io(mdev, sector);
}
- remote = remote && drbd_should_do_remote(mdev->state);
- send_oos = rw == WRITE && drbd_should_send_oos(mdev->state);
+ s = mdev->state;
+ remote = remote && drbd_should_do_remote(s);
+ send_oos = rw == WRITE && drbd_should_send_oos(s);
D_ASSERT(!(remote && send_oos));
if (!(local || remote) && !is_susp(mdev->state)) {
@@ -867,7 +871,7 @@ allocate_barrier:
if (is_susp(mdev->state)) {
/* If we got suspended, use the retry mechanism of
- generic_make_request() to restart processing of this
+ drbd_make_request() to restart processing of this
bio. In the next call to drbd_make_request
we sleep in inc_ap_bio() */
ret = 1;
@@ -1091,7 +1095,6 @@ void drbd_make_request(struct request_queue *q, struct bio *bio)
*/
D_ASSERT(bio->bi_size > 0);
D_ASSERT((bio->bi_size & 0x1ff) == 0);
- D_ASSERT(bio->bi_idx == 0);
/* to make some things easier, force alignment of requests within the
* granularity of our hash tables */
@@ -1099,8 +1102,9 @@ void drbd_make_request(struct request_queue *q, struct bio *bio)
e_enr = (bio->bi_sector+(bio->bi_size>>9)-1) >> HT_SHIFT;
if (likely(s_enr == e_enr)) {
- inc_ap_bio(mdev, 1);
- drbd_make_request_common(mdev, bio, start_time);
+ do {
+ inc_ap_bio(mdev, 1);
+ } while (drbd_make_request_common(mdev, bio, start_time));
return;
}
@@ -1196,36 +1200,66 @@ void request_timer_fn(unsigned long data)
struct drbd_conf *mdev = (struct drbd_conf *) data;
struct drbd_request *req; /* oldest request */
struct list_head *le;
- unsigned long et = 0; /* effective timeout = ko_count * timeout */
+ unsigned long ent = 0, dt = 0, et, nt; /* effective timeout = ko_count * timeout */
+ unsigned long now;
if (get_net_conf(mdev)) {
- et = mdev->net_conf->timeout*HZ/10 * mdev->net_conf->ko_count;
+ if (mdev->state.conn >= C_WF_REPORT_PARAMS)
+ ent = mdev->net_conf->timeout*HZ/10
+ * mdev->net_conf->ko_count;
put_net_conf(mdev);
}
- if (!et || mdev->state.conn < C_WF_REPORT_PARAMS)
+ if (get_ldev(mdev)) { /* implicit state.disk >= D_INCONSISTENT */
+ dt = mdev->ldev->dc.disk_timeout * HZ / 10;
+ put_ldev(mdev);
+ }
+ et = min_not_zero(dt, ent);
+
+ if (!et)
return; /* Recurring timer stopped */
+ now = jiffies;
+
spin_lock_irq(&mdev->req_lock);
le = &mdev->oldest_tle->requests;
if (list_empty(le)) {
spin_unlock_irq(&mdev->req_lock);
- mod_timer(&mdev->request_timer, jiffies + et);
+ mod_timer(&mdev->request_timer, now + et);
return;
}
le = le->prev;
req = list_entry(le, struct drbd_request, tl_requests);
- if (time_is_before_eq_jiffies(req->start_time + et)) {
- if (req->rq_state & RQ_NET_PENDING) {
- dev_warn(DEV, "Remote failed to finish a request within ko-count * timeout\n");
- _drbd_set_state(_NS(mdev, conn, C_TIMEOUT), CS_VERBOSE, NULL);
- } else {
- dev_warn(DEV, "Local backing block device frozen?\n");
- mod_timer(&mdev->request_timer, jiffies + et);
- }
- } else {
- mod_timer(&mdev->request_timer, req->start_time + et);
- }
+ /* The request is considered timed out, if
+ * - we have some effective timeout from the configuration,
+ * with above state restrictions applied,
+ * - the oldest request is waiting for a response from the network
+ * resp. the local disk,
+ * - the oldest request is in fact older than the effective timeout,
+ * - the connection was established (resp. disk was attached)
+ * for longer than the timeout already.
+ * Note that for 32bit jiffies and very stable connections/disks,
+ * we may have a wrap around, which is catched by
+ * !time_in_range(now, last_..._jif, last_..._jif + timeout).
+ *
+ * Side effect: once per 32bit wrap-around interval, which means every
+ * ~198 days with 250 HZ, we have a window where the timeout would need
+ * to expire twice (worst case) to become effective. Good enough.
+ */
+ if (ent && req->rq_state & RQ_NET_PENDING &&
+ time_after(now, req->start_time + ent) &&
+ !time_in_range(now, mdev->last_reconnect_jif, mdev->last_reconnect_jif + ent)) {
+ dev_warn(DEV, "Remote failed to finish a request within ko-count * timeout\n");
+ _drbd_set_state(_NS(mdev, conn, C_TIMEOUT), CS_VERBOSE | CS_HARD, NULL);
+ }
+ if (dt && req->rq_state & RQ_LOCAL_PENDING &&
+ 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);
+ }
+ nt = (time_after(now, req->start_time + et) ? now : req->start_time) + et;
spin_unlock_irq(&mdev->req_lock);
+ mod_timer(&mdev->request_timer, nt);
}
diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h
index 68a234a5fdc5..3d2111919486 100644
--- a/drivers/block/drbd/drbd_req.h
+++ b/drivers/block/drbd/drbd_req.h
@@ -105,6 +105,7 @@ enum drbd_req_event {
read_completed_with_error,
read_ahead_completed_with_error,
write_completed_with_error,
+ abort_disk_io,
completed_ok,
resend,
fail_frozen_disk_io,
@@ -118,18 +119,21 @@ enum drbd_req_event {
* same time, so we should hold the request lock anyways.
*/
enum drbd_req_state_bits {
- /* 210
- * 000: no local possible
- * 001: to be submitted
+ /* 3210
+ * 0000: no local possible
+ * 0001: to be submitted
* UNUSED, we could map: 011: submitted, completion still pending
- * 110: completed ok
- * 010: completed with error
+ * 0110: completed ok
+ * 0010: completed with error
+ * 1001: Aborted (before completion)
+ * 1x10: Aborted and completed -> free
*/
__RQ_LOCAL_PENDING,
__RQ_LOCAL_COMPLETED,
__RQ_LOCAL_OK,
+ __RQ_LOCAL_ABORTED,
- /* 76543
+ /* 87654
* 00000: no network possible
* 00001: to be send
* 00011: to be send, on worker queue
@@ -199,8 +203,9 @@ enum drbd_req_state_bits {
#define RQ_LOCAL_PENDING (1UL << __RQ_LOCAL_PENDING)
#define RQ_LOCAL_COMPLETED (1UL << __RQ_LOCAL_COMPLETED)
#define RQ_LOCAL_OK (1UL << __RQ_LOCAL_OK)
+#define RQ_LOCAL_ABORTED (1UL << __RQ_LOCAL_ABORTED)
-#define RQ_LOCAL_MASK ((RQ_LOCAL_OK << 1)-1) /* 0x07 */
+#define RQ_LOCAL_MASK ((RQ_LOCAL_ABORTED << 1)-1)
#define RQ_NET_PENDING (1UL << __RQ_NET_PENDING)
#define RQ_NET_QUEUED (1UL << __RQ_NET_QUEUED)
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index 4d3e6f6213ba..620c70ff2231 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -70,11 +70,29 @@ rwlock_t global_state_lock;
void drbd_md_io_complete(struct bio *bio, int error)
{
struct drbd_md_io *md_io;
+ struct drbd_conf *mdev;
md_io = (struct drbd_md_io *)bio->bi_private;
+ mdev = container_of(md_io, struct drbd_conf, md_io);
+
md_io->error = error;
- complete(&md_io->event);
+ /* 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.
+ * If this io completion runs after that timeout expired, this
+ * drbd_md_put_buffer() may allow us to finally try and re-attach.
+ * During normal operation, this only puts that extra reference
+ * down to 1 again.
+ * 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.
+ */
+ drbd_md_put_buffer(mdev);
+ md_io->done = 1;
+ wake_up(&mdev->misc_wait);
+ bio_put(bio);
+ put_ldev(mdev);
}
/* reads on behalf of the partner,
@@ -226,6 +244,7 @@ void drbd_endio_pri(struct bio *bio, int error)
spin_lock_irqsave(&mdev->req_lock, flags);
__req_mod(req, what, &m);
spin_unlock_irqrestore(&mdev->req_lock, flags);
+ put_ldev(mdev);
if (m.bio)
complete_master_bio(mdev, &m);
@@ -290,7 +309,7 @@ void drbd_csum_bio(struct drbd_conf *mdev, struct crypto_hash *tfm, struct bio *
sg_init_table(&sg, 1);
crypto_hash_init(&desc);
- __bio_for_each_segment(bvec, bio, i, 0) {
+ bio_for_each_segment(bvec, bio, i) {
sg_set_page(&sg, bvec->bv_page, bvec->bv_len, bvec->bv_offset);
crypto_hash_update(&desc, &sg, sg.length);
}
@@ -728,7 +747,7 @@ int w_start_resync(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
}
drbd_start_resync(mdev, C_SYNC_SOURCE);
- clear_bit(AHEAD_TO_SYNC_SOURCE, &mdev->current_epoch->flags);
+ clear_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags);
return 1;
}
@@ -1519,14 +1538,14 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side)
}
drbd_state_lock(mdev);
-
+ write_lock_irq(&global_state_lock);
if (!get_ldev_if_state(mdev, D_NEGOTIATING)) {
+ write_unlock_irq(&global_state_lock);
drbd_state_unlock(mdev);
return;
}
- write_lock_irq(&global_state_lock);
- ns = mdev->state;
+ ns.i = mdev->state.i;
ns.aftr_isp = !_drbd_may_sync_now(mdev);
diff --git a/drivers/block/floppy.c b/drivers/block/floppy.c
index b0b00d70c166..cce7df367b79 100644
--- a/drivers/block/floppy.c
+++ b/drivers/block/floppy.c
@@ -551,7 +551,7 @@ static void floppy_ready(void);
static void floppy_start(void);
static void process_fd_request(void);
static void recalibrate_floppy(void);
-static void floppy_shutdown(unsigned long);
+static void floppy_shutdown(struct work_struct *);
static int floppy_request_regions(int);
static void floppy_release_regions(int);
@@ -588,6 +588,8 @@ static int buffer_max = -1;
static struct floppy_fdc_state fdc_state[N_FDC];
static int fdc; /* current fdc */
+static struct workqueue_struct *floppy_wq;
+
static struct floppy_struct *_floppy = floppy_type;
static unsigned char current_drive;
static long current_count_sectors;
@@ -629,16 +631,15 @@ static inline void set_debugt(void) { }
static inline void debugt(const char *func, const char *msg) { }
#endif /* DEBUGT */
-typedef void (*timeout_fn)(unsigned long);
-static DEFINE_TIMER(fd_timeout, floppy_shutdown, 0, 0);
+static DECLARE_DELAYED_WORK(fd_timeout, floppy_shutdown);
static const char *timeout_message;
static void is_alive(const char *func, const char *message)
{
/* this routine checks whether the floppy driver is "alive" */
if (test_bit(0, &fdc_busy) && command_status < 2 &&
- !timer_pending(&fd_timeout)) {
+ !delayed_work_pending(&fd_timeout)) {
DPRINT("%s: timeout handler died. %s\n", func, message);
}
}
@@ -666,15 +667,18 @@ static int output_log_pos;
static void __reschedule_timeout(int drive, const char *message)
{
+ unsigned long delay;
+
if (drive == current_reqD)
drive = current_drive;
- del_timer(&fd_timeout);
+
if (drive < 0 || drive >= N_DRIVE) {
- fd_timeout.expires = jiffies + 20UL * HZ;
+ delay = 20UL * HZ;
drive = 0;
} else
- fd_timeout.expires = jiffies + UDP->timeout;
- add_timer(&fd_timeout);
+ delay = UDP->timeout;
+
+ queue_delayed_work(floppy_wq, &fd_timeout, delay);
if (UDP->flags & FD_DEBUG)
DPRINT("reschedule timeout %s\n", message);
timeout_message = message;
@@ -872,7 +876,7 @@ static int lock_fdc(int drive, bool interruptible)
command_status = FD_COMMAND_NONE;
- __reschedule_timeout(drive, "lock fdc");
+ reschedule_timeout(drive, "lock fdc");
set_fdc(drive);
return 0;
}
@@ -880,23 +884,15 @@ static int lock_fdc(int drive, bool interruptible)
/* unlocks the driver */
static void unlock_fdc(void)
{
- unsigned long flags;
-
- raw_cmd = NULL;
if (!test_bit(0, &fdc_busy))
DPRINT("FDC access conflict!\n");
- if (do_floppy)
- DPRINT("device interrupt still active at FDC release: %pf!\n",
- do_floppy);
+ raw_cmd = NULL;
command_status = FD_COMMAND_NONE;
- spin_lock_irqsave(&floppy_lock, flags);
- del_timer(&fd_timeout);
+ __cancel_delayed_work(&fd_timeout);
+ do_floppy = NULL;
cont = NULL;
clear_bit(0, &fdc_busy);
- if (current_req || set_next_request())
- do_fd_request(current_req->q);
- spin_unlock_irqrestore(&floppy_lock, flags);
wake_up(&fdc_wait);
}
@@ -968,26 +964,24 @@ static DECLARE_WORK(floppy_work, NULL);
static void schedule_bh(void (*handler)(void))
{
+ WARN_ON(work_pending(&floppy_work));
+
PREPARE_WORK(&floppy_work, (work_func_t)handler);
- schedule_work(&floppy_work);
+ queue_work(floppy_wq, &floppy_work);
}
-static DEFINE_TIMER(fd_timer, NULL, 0, 0);
+static DECLARE_DELAYED_WORK(fd_timer, NULL);
static void cancel_activity(void)
{
- unsigned long flags;
-
- spin_lock_irqsave(&floppy_lock, flags);
do_floppy = NULL;
- PREPARE_WORK(&floppy_work, (work_func_t)empty);
- del_timer(&fd_timer);
- spin_unlock_irqrestore(&floppy_lock, flags);
+ cancel_delayed_work_sync(&fd_timer);
+ cancel_work_sync(&floppy_work);
}
/* this function makes sure that the disk stays in the drive during the
* transfer */
-static void fd_watchdog(void)
+static void fd_watchdog(struct work_struct *arg)
{
debug_dcl(DP->flags, "calling disk change from watchdog\n");
@@ -997,21 +991,20 @@ static void fd_watchdog(void)
cont->done(0);
reset_fdc();
} else {
- del_timer(&fd_timer);
- fd_timer.function = (timeout_fn)fd_watchdog;
- fd_timer.expires = jiffies + HZ / 10;
- add_timer(&fd_timer);
+ cancel_delayed_work(&fd_timer);
+ PREPARE_DELAYED_WORK(&fd_timer, fd_watchdog);
+ queue_delayed_work(floppy_wq, &fd_timer, HZ / 10);
}
}
static void main_command_interrupt(void)
{
- del_timer(&fd_timer);
+ cancel_delayed_work(&fd_timer);
cont->interrupt();
}
/* waits for a delay (spinup or select) to pass */
-static int fd_wait_for_completion(unsigned long delay, timeout_fn function)
+static int fd_wait_for_completion(unsigned long expires, work_func_t function)
{
if (FDCS->reset) {
reset_fdc(); /* do the reset during sleep to win time
@@ -1020,11 +1013,10 @@ static int fd_wait_for_completion(unsigned long delay, timeout_fn function)
return 1;
}
- if (time_before(jiffies, delay)) {
- del_timer(&fd_timer);
- fd_timer.function = function;
- fd_timer.expires = delay;
- add_timer(&fd_timer);
+ if (time_before(jiffies, expires)) {
+ cancel_delayed_work(&fd_timer);
+ PREPARE_DELAYED_WORK(&fd_timer, function);
+ queue_delayed_work(floppy_wq, &fd_timer, expires - jiffies);
return 1;
}
return 0;
@@ -1342,7 +1334,7 @@ static int fdc_dtr(void)
*/
FDCS->dtr = raw_cmd->rate & 3;
return fd_wait_for_completion(jiffies + 2UL * HZ / 100,
- (timeout_fn)floppy_ready);
+ (work_func_t)floppy_ready);
} /* fdc_dtr */
static void tell_sector(void)
@@ -1447,7 +1439,7 @@ static void setup_rw_floppy(void)
int flags;
int dflags;
unsigned long ready_date;
- timeout_fn function;
+ work_func_t function;
flags = raw_cmd->flags;
if (flags & (FD_RAW_READ | FD_RAW_WRITE))
@@ -1461,9 +1453,9 @@ static void setup_rw_floppy(void)
*/
if (time_after(ready_date, jiffies + DP->select_delay)) {
ready_date -= DP->select_delay;
- function = (timeout_fn)floppy_start;
+ function = (work_func_t)floppy_start;
} else
- function = (timeout_fn)setup_rw_floppy;
+ function = (work_func_t)setup_rw_floppy;
/* wait until the floppy is spinning fast enough */
if (fd_wait_for_completion(ready_date, function))
@@ -1493,7 +1485,7 @@ static void setup_rw_floppy(void)
inr = result();
cont->interrupt();
} else if (flags & FD_RAW_NEED_DISK)
- fd_watchdog();
+ fd_watchdog(NULL);
}
static int blind_seek;
@@ -1802,20 +1794,22 @@ static void show_floppy(void)
pr_info("do_floppy=%pf\n", do_floppy);
if (work_pending(&floppy_work))
pr_info("floppy_work.func=%pf\n", floppy_work.func);
- if (timer_pending(&fd_timer))
- pr_info("fd_timer.function=%pf\n", fd_timer.function);
- if (timer_pending(&fd_timeout)) {
- pr_info("timer_function=%pf\n", fd_timeout.function);
- pr_info("expires=%lu\n", fd_timeout.expires - jiffies);
- pr_info("now=%lu\n", jiffies);
- }
+ if (delayed_work_pending(&fd_timer))
+ pr_info("delayed work.function=%p expires=%ld\n",
+ fd_timer.work.func,
+ fd_timer.timer.expires - jiffies);
+ if (delayed_work_pending(&fd_timeout))
+ pr_info("timer_function=%p expires=%ld\n",
+ fd_timeout.work.func,
+ fd_timeout.timer.expires - jiffies);
+
pr_info("cont=%p\n", cont);
pr_info("current_req=%p\n", current_req);
pr_info("command_status=%d\n", command_status);
pr_info("\n");
}
-static void floppy_shutdown(unsigned long data)
+static void floppy_shutdown(struct work_struct *arg)
{
unsigned long flags;
@@ -1868,7 +1862,7 @@ static int start_motor(void (*function)(void))
/* wait_for_completion also schedules reset if needed. */
return fd_wait_for_completion(DRS->select_date + DP->select_delay,
- (timeout_fn)function);
+ (work_func_t)function);
}
static void floppy_ready(void)
@@ -2821,7 +2815,6 @@ do_request:
spin_lock_irq(&floppy_lock);
pending = set_next_request();
spin_unlock_irq(&floppy_lock);
-
if (!pending) {
do_floppy = NULL;
unlock_fdc();
@@ -2898,13 +2891,15 @@ static void do_fd_request(struct request_queue *q)
current_req->cmd_flags))
return;
- if (test_bit(0, &fdc_busy)) {
+ if (test_and_set_bit(0, &fdc_busy)) {
/* fdc busy, this new request will be treated when the
current one is done */
is_alive(__func__, "old request running");
return;
}
- lock_fdc(MAXTIMEOUT, false);
+ command_status = FD_COMMAND_NONE;
+ __reschedule_timeout(MAXTIMEOUT, "fd_request");
+ set_fdc(0);
process_fd_request();
is_alive(__func__, "");
}
@@ -3612,9 +3607,7 @@ static int floppy_release(struct gendisk *disk, fmode_t mode)
mutex_lock(&floppy_mutex);
mutex_lock(&open_lock);
- if (UDRS->fd_ref < 0)
- UDRS->fd_ref = 0;
- else if (!UDRS->fd_ref--) {
+ if (!UDRS->fd_ref--) {
DPRINT("floppy_release with fd_ref == 0");
UDRS->fd_ref = 0;
}
@@ -3650,13 +3643,7 @@ static int floppy_open(struct block_device *bdev, fmode_t mode)
set_bit(FD_VERIFY_BIT, &UDRS->flags);
}
- if (UDRS->fd_ref == -1 || (UDRS->fd_ref && (mode & FMODE_EXCL)))
- goto out2;
-
- if (mode & FMODE_EXCL)
- UDRS->fd_ref = -1;
- else
- UDRS->fd_ref++;
+ UDRS->fd_ref++;
opened_bdev[drive] = bdev;
@@ -3719,10 +3706,8 @@ static int floppy_open(struct block_device *bdev, fmode_t mode)
mutex_unlock(&floppy_mutex);
return 0;
out:
- if (UDRS->fd_ref < 0)
- UDRS->fd_ref = 0;
- else
- UDRS->fd_ref--;
+ UDRS->fd_ref--;
+
if (!UDRS->fd_ref)
opened_bdev[drive] = NULL;
out2:
@@ -4159,10 +4144,16 @@ static int __init floppy_init(void)
goto out_put_disk;
}
+ floppy_wq = alloc_ordered_workqueue("floppy", 0);
+ if (!floppy_wq) {
+ err = -ENOMEM;
+ goto out_put_disk;
+ }
+
disks[dr]->queue = blk_init_queue(do_fd_request, &floppy_lock);
if (!disks[dr]->queue) {
err = -ENOMEM;
- goto out_put_disk;
+ goto out_destroy_workq;
}
blk_queue_max_hw_sectors(disks[dr]->queue, 64);
@@ -4213,7 +4204,7 @@ static int __init floppy_init(void)
use_virtual_dma = can_use_virtual_dma & 1;
fdc_state[0].address = FDC1;
if (fdc_state[0].address == -1) {
- del_timer_sync(&fd_timeout);
+ cancel_delayed_work(&fd_timeout);
err = -ENODEV;
goto out_unreg_region;
}
@@ -4224,7 +4215,7 @@ static int __init floppy_init(void)
fdc = 0; /* reset fdc in case of unexpected interrupt */
err = floppy_grab_irq_and_dma();
if (err) {
- del_timer_sync(&fd_timeout);
+ cancel_delayed_work(&fd_timeout);
err = -EBUSY;
goto out_unreg_region;
}
@@ -4281,13 +4272,13 @@ static int __init floppy_init(void)
user_reset_fdc(-1, FD_RESET_ALWAYS, false);
}
fdc = 0;
- del_timer_sync(&fd_timeout);
+ cancel_delayed_work(&fd_timeout);
current_drive = 0;
initialized = true;
if (have_no_fdc) {
DPRINT("no floppy controllers found\n");
err = have_no_fdc;
- goto out_flush_work;
+ goto out_release_dma;
}
for (drive = 0; drive < N_DRIVE; drive++) {
@@ -4302,7 +4293,7 @@ static int __init floppy_init(void)
err = platform_device_register(&floppy_device[drive]);
if (err)
- goto out_flush_work;
+ goto out_release_dma;
err = device_create_file(&floppy_device[drive].dev,
&dev_attr_cmos);
@@ -4320,13 +4311,14 @@ static int __init floppy_init(void)
out_unreg_platform_dev:
platform_device_unregister(&floppy_device[drive]);
-out_flush_work:
- flush_work_sync(&floppy_work);
+out_release_dma:
if (atomic_read(&usage_count))
floppy_release_irq_and_dma();
out_unreg_region:
blk_unregister_region(MKDEV(FLOPPY_MAJOR, 0), 256);
platform_driver_unregister(&floppy_driver);
+out_destroy_workq:
+ destroy_workqueue(floppy_wq);
out_unreg_blkdev:
unregister_blkdev(FLOPPY_MAJOR, "fd");
out_put_disk:
@@ -4397,7 +4389,7 @@ static int floppy_grab_irq_and_dma(void)
* We might have scheduled a free_irq(), wait it to
* drain first:
*/
- flush_work_sync(&floppy_work);
+ flush_workqueue(floppy_wq);
if (fd_request_irq()) {
DPRINT("Unable to grab IRQ%d for the floppy driver\n",
@@ -4488,9 +4480,9 @@ static void floppy_release_irq_and_dma(void)
pr_info("motor off timer %d still active\n", drive);
#endif
- if (timer_pending(&fd_timeout))
+ if (delayed_work_pending(&fd_timeout))
pr_info("floppy timer still active:%s\n", timeout_message);
- if (timer_pending(&fd_timer))
+ if (delayed_work_pending(&fd_timer))
pr_info("auxiliary floppy timer still active\n");
if (work_pending(&floppy_work))
pr_info("work still pending\n");
@@ -4560,8 +4552,9 @@ static void __exit floppy_module_exit(void)
put_disk(disks[drive]);
}
- del_timer_sync(&fd_timeout);
- del_timer_sync(&fd_timer);
+ cancel_delayed_work_sync(&fd_timeout);
+ cancel_delayed_work_sync(&fd_timer);
+ destroy_workqueue(floppy_wq);
if (atomic_read(&usage_count))
floppy_release_irq_and_dma();
diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c
index 304000c3d433..264bc77dcb91 100644
--- a/drivers/block/mtip32xx/mtip32xx.c
+++ b/drivers/block/mtip32xx/mtip32xx.c
@@ -294,18 +294,16 @@ static int hba_reset_nosleep(struct driver_data *dd)
*/
static inline void mtip_issue_ncq_command(struct mtip_port *port, int tag)
{
- unsigned long flags = 0;
-
atomic_set(&port->commands[tag].active, 1);
- spin_lock_irqsave(&port->cmd_issue_lock, flags);
+ spin_lock(&port->cmd_issue_lock);
writel((1 << MTIP_TAG_BIT(tag)),
port->s_active[MTIP_TAG_INDEX(tag)]);
writel((1 << MTIP_TAG_BIT(tag)),
port->cmd_issue[MTIP_TAG_INDEX(tag)]);
- spin_unlock_irqrestore(&port->cmd_issue_lock, flags);
+ spin_unlock(&port->cmd_issue_lock);
/* Set the command's timeout value.*/
port->commands[tag].comp_time = jiffies + msecs_to_jiffies(
@@ -436,8 +434,7 @@ static void mtip_init_port(struct mtip_port *port)
writel(0xFFFFFFFF, port->completed[i]);
/* Clear any pending interrupts for this port */
- writel(readl(port->dd->mmio + PORT_IRQ_STAT),
- port->dd->mmio + PORT_IRQ_STAT);
+ writel(readl(port->mmio + PORT_IRQ_STAT), port->mmio + PORT_IRQ_STAT);
/* Clear any pending interrupts on the HBA. */
writel(readl(port->dd->mmio + HOST_IRQ_STAT),
@@ -782,13 +779,24 @@ static void mtip_handle_tfe(struct driver_data *dd)
/* Stop the timer to prevent command timeouts. */
del_timer(&port->cmd_timer);
+ set_bit(MTIP_PF_EH_ACTIVE_BIT, &port->flags);
+
+ if (test_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags) &&
+ test_bit(MTIP_TAG_INTERNAL, port->allocated)) {
+ cmd = &port->commands[MTIP_TAG_INTERNAL];
+ dbg_printk(MTIP_DRV_NAME " TFE for the internal command\n");
+
+ atomic_inc(&cmd->active); /* active > 1 indicates error */
+ if (cmd->comp_data && cmd->comp_func) {
+ cmd->comp_func(port, MTIP_TAG_INTERNAL,
+ cmd->comp_data, PORT_IRQ_TF_ERR);
+ }
+ goto handle_tfe_exit;
+ }
/* clear the tag accumulator */
memset(tagaccum, 0, SLOTBITS_IN_LONGS * sizeof(long));
- /* Set eh_active */
- set_bit(MTIP_PF_EH_ACTIVE_BIT, &port->flags);
-
/* Loop through all the groups */
for (group = 0; group < dd->slot_groups; group++) {
completed = readl(port->completed[group]);
@@ -940,6 +948,7 @@ static void mtip_handle_tfe(struct driver_data *dd)
}
print_tags(dd, "reissued (TFE)", tagaccum, cmd_cnt);
+handle_tfe_exit:
/* clear eh_active */
clear_bit(MTIP_PF_EH_ACTIVE_BIT, &port->flags);
wake_up_interruptible(&port->svc_wait);
@@ -961,6 +970,8 @@ static inline void mtip_process_sdbf(struct driver_data *dd)
/* walk all bits in all slot groups */
for (group = 0; group < dd->slot_groups; group++) {
completed = readl(port->completed[group]);
+ if (!completed)
+ continue;
/* clear completed status register in the hardware.*/
writel(completed, port->completed[group]);
@@ -1329,22 +1340,6 @@ static int mtip_exec_internal_command(struct mtip_port *port,
}
rv = -EAGAIN;
}
-
- if (readl(port->cmd_issue[MTIP_TAG_INTERNAL])
- & (1 << MTIP_TAG_INTERNAL)) {
- dev_warn(&port->dd->pdev->dev,
- "Retiring internal command but CI is 1.\n");
- if (test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
- &port->dd->dd_flag)) {
- hba_reset_nosleep(port->dd);
- rv = -ENXIO;
- } else {
- mtip_restart_port(port);
- rv = -EAGAIN;
- }
- goto exec_ic_exit;
- }
-
} else {
/* Spin for <timeout> checking if command still outstanding */
timeout = jiffies + msecs_to_jiffies(timeout);
@@ -1361,21 +1356,25 @@ static int mtip_exec_internal_command(struct mtip_port *port,
rv = -ENXIO;
goto exec_ic_exit;
}
+ if (readl(port->mmio + PORT_IRQ_STAT) & PORT_IRQ_ERR) {
+ atomic_inc(&int_cmd->active); /* error */
+ break;
+ }
}
+ }
- if (readl(port->cmd_issue[MTIP_TAG_INTERNAL])
+ if (atomic_read(&int_cmd->active) > 1) {
+ dev_err(&port->dd->pdev->dev,
+ "Internal command [%02X] failed\n", fis->command);
+ rv = -EIO;
+ }
+ if (readl(port->cmd_issue[MTIP_TAG_INTERNAL])
& (1 << MTIP_TAG_INTERNAL)) {
- dev_err(&port->dd->pdev->dev,
- "Internal command did not complete [atomic]\n");
+ rv = -ENXIO;
+ if (!test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
+ &port->dd->dd_flag)) {
+ mtip_restart_port(port);
rv = -EAGAIN;
- if (test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
- &port->dd->dd_flag)) {
- hba_reset_nosleep(port->dd);
- rv = -ENXIO;
- } else {
- mtip_restart_port(port);
- rv = -EAGAIN;
- }
}
}
exec_ic_exit:
@@ -1893,13 +1892,33 @@ static int exec_drive_command(struct mtip_port *port, u8 *command,
void __user *user_buffer)
{
struct host_to_dev_fis fis;
- struct host_to_dev_fis *reply = (port->rxfis + RX_FIS_D2H_REG);
+ struct host_to_dev_fis *reply;
+ u8 *buf = NULL;
+ dma_addr_t dma_addr = 0;
+ int rv = 0, xfer_sz = command[3];
+
+ if (xfer_sz) {
+ if (user_buffer)
+ return -EFAULT;
+
+ buf = dmam_alloc_coherent(&port->dd->pdev->dev,
+ ATA_SECT_SIZE * xfer_sz,
+ &dma_addr,
+ GFP_KERNEL);
+ if (!buf) {
+ dev_err(&port->dd->pdev->dev,
+ "Memory allocation failed (%d bytes)\n",
+ ATA_SECT_SIZE * xfer_sz);
+ return -ENOMEM;
+ }
+ memset(buf, 0, ATA_SECT_SIZE * xfer_sz);
+ }
/* Build the FIS. */
memset(&fis, 0, sizeof(struct host_to_dev_fis));
- fis.type = 0x27;
- fis.opts = 1 << 7;
- fis.command = command[0];
+ fis.type = 0x27;
+ fis.opts = 1 << 7;
+ fis.command = command[0];
fis.features = command[2];
fis.sect_count = command[3];
if (fis.command == ATA_CMD_SMART) {
@@ -1908,6 +1927,11 @@ static int exec_drive_command(struct mtip_port *port, u8 *command,
fis.cyl_hi = 0xC2;
}
+ if (xfer_sz)
+ reply = (port->rxfis + RX_FIS_PIO_SETUP);
+ else
+ reply = (port->rxfis + RX_FIS_D2H_REG);
+
dbg_printk(MTIP_DRV_NAME
" %s: User Command: cmd %x, sect %x, "
"feat %x, sectcnt %x\n",
@@ -1917,43 +1941,46 @@ static int exec_drive_command(struct mtip_port *port, u8 *command,
command[2],
command[3]);
- memset(port->sector_buffer, 0x00, ATA_SECT_SIZE);
-
/* Execute the command. */
if (mtip_exec_internal_command(port,
&fis,
5,
- port->sector_buffer_dma,
- (command[3] != 0) ? ATA_SECT_SIZE : 0,
+ (xfer_sz ? dma_addr : 0),
+ (xfer_sz ? ATA_SECT_SIZE * xfer_sz : 0),
0,
GFP_KERNEL,
MTIP_IOCTL_COMMAND_TIMEOUT_MS)
< 0) {
- return -1;
+ rv = -EFAULT;
+ goto exit_drive_command;
}
/* Collect the completion status. */
command[0] = reply->command; /* Status*/
command[1] = reply->features; /* Error*/
- command[2] = command[3];
+ command[2] = reply->sect_count;
dbg_printk(MTIP_DRV_NAME
" %s: Completion Status: stat %x, "
- "err %x, cmd %x\n",
+ "err %x, nsect %x\n",
__func__,
command[0],
command[1],
command[2]);
- if (user_buffer && command[3]) {
+ if (xfer_sz) {
if (copy_to_user(user_buffer,
- port->sector_buffer,
+ buf,
ATA_SECT_SIZE * command[3])) {
- return -EFAULT;
+ rv = -EFAULT;
+ goto exit_drive_command;
}
}
-
- return 0;
+exit_drive_command:
+ if (buf)
+ dmam_free_coherent(&port->dd->pdev->dev,
+ ATA_SECT_SIZE * xfer_sz, buf, dma_addr);
+ return rv;
}
/*
@@ -2003,6 +2030,32 @@ static unsigned int implicit_sector(unsigned char command,
return rv;
}
+static void mtip_set_timeout(struct host_to_dev_fis *fis, unsigned int *timeout)
+{
+ switch (fis->command) {
+ case ATA_CMD_DOWNLOAD_MICRO:
+ *timeout = 120000; /* 2 minutes */
+ break;
+ case ATA_CMD_SEC_ERASE_UNIT:
+ case 0xFC:
+ *timeout = 240000; /* 4 minutes */
+ break;
+ case ATA_CMD_STANDBYNOW1:
+ *timeout = 10000; /* 10 seconds */
+ break;
+ case 0xF7:
+ case 0xFA:
+ *timeout = 60000; /* 60 seconds */
+ break;
+ case ATA_CMD_SMART:
+ *timeout = 15000; /* 15 seconds */
+ break;
+ default:
+ *timeout = MTIP_IOCTL_COMMAND_TIMEOUT_MS;
+ break;
+ }
+}
+
/*
* Executes a taskfile
* See ide_taskfile_ioctl() for derivation
@@ -2023,7 +2076,7 @@ static int exec_drive_taskfile(struct driver_data *dd,
unsigned int taskin = 0;
unsigned int taskout = 0;
u8 nsect = 0;
- unsigned int timeout = MTIP_IOCTL_COMMAND_TIMEOUT_MS;
+ unsigned int timeout;
unsigned int force_single_sector;
unsigned int transfer_size;
unsigned long task_file_data;
@@ -2153,32 +2206,7 @@ static int exec_drive_taskfile(struct driver_data *dd,
fis.lba_hi,
fis.device);
- switch (fis.command) {
- case ATA_CMD_DOWNLOAD_MICRO:
- /* Change timeout for Download Microcode to 2 minutes */
- timeout = 120000;
- break;
- case ATA_CMD_SEC_ERASE_UNIT:
- /* Change timeout for Security Erase Unit to 4 minutes.*/
- timeout = 240000;
- break;
- case ATA_CMD_STANDBYNOW1:
- /* Change timeout for standby immediate to 10 seconds.*/
- timeout = 10000;
- break;
- case 0xF7:
- case 0xFA:
- /* Change timeout for vendor unique command to 10 secs */
- timeout = 10000;
- break;
- case ATA_CMD_SMART:
- /* Change timeout for vendor unique command to 15 secs */
- timeout = 15000;
- break;
- default:
- timeout = MTIP_IOCTL_COMMAND_TIMEOUT_MS;
- break;
- }
+ mtip_set_timeout(&fis, &timeout);
/* Determine the correct transfer size.*/
if (force_single_sector)
@@ -2295,13 +2323,12 @@ static int mtip_hw_ioctl(struct driver_data *dd, unsigned int cmd,
{
switch (cmd) {
case HDIO_GET_IDENTITY:
- if (mtip_get_identify(dd->port, (void __user *) arg) < 0) {
- dev_warn(&dd->pdev->dev,
- "Unable to read identity\n");
- return -EIO;
- }
-
+ {
+ if (copy_to_user((void __user *)arg, dd->port->identify,
+ sizeof(u16) * ATA_ID_WORDS))
+ return -EFAULT;
break;
+ }
case HDIO_DRIVE_CMD:
{
u8 drive_command[4];
@@ -2537,40 +2564,58 @@ static ssize_t mtip_hw_show_registers(struct device *dev,
int size = 0;
int n;
- size += sprintf(&buf[size], "S ACTive:\n");
+ size += sprintf(&buf[size], "Hardware\n--------\n");
+ size += sprintf(&buf[size], "S ACTive : [ 0x");
- for (n = 0; n < dd->slot_groups; n++)
- size += sprintf(&buf[size], "0x%08x\n",
+ for (n = dd->slot_groups-1; n >= 0; n--)
+ size += sprintf(&buf[size], "%08X ",
readl(dd->port->s_active[n]));
- size += sprintf(&buf[size], "Command Issue:\n");
+ size += sprintf(&buf[size], "]\n");
+ size += sprintf(&buf[size], "Command Issue : [ 0x");
- for (n = 0; n < dd->slot_groups; n++)
- size += sprintf(&buf[size], "0x%08x\n",
+ for (n = dd->slot_groups-1; n >= 0; n--)
+ size += sprintf(&buf[size], "%08X ",
readl(dd->port->cmd_issue[n]));
- size += sprintf(&buf[size], "Allocated:\n");
+ size += sprintf(&buf[size], "]\n");
+ size += sprintf(&buf[size], "Completed : [ 0x");
+
+ for (n = dd->slot_groups-1; n >= 0; n--)
+ size += sprintf(&buf[size], "%08X ",
+ readl(dd->port->completed[n]));
+
+ size += sprintf(&buf[size], "]\n");
+ size += sprintf(&buf[size], "PORT IRQ STAT : [ 0x%08X ]\n",
+ readl(dd->port->mmio + PORT_IRQ_STAT));
+ size += sprintf(&buf[size], "HOST IRQ STAT : [ 0x%08X ]\n",
+ readl(dd->mmio + HOST_IRQ_STAT));
+ size += sprintf(&buf[size], "\n");
- for (n = 0; n < dd->slot_groups; n++) {
+ size += sprintf(&buf[size], "Local\n-----\n");
+ size += sprintf(&buf[size], "Allocated : [ 0x");
+
+ for (n = dd->slot_groups-1; n >= 0; n--) {
if (sizeof(long) > sizeof(u32))
group_allocated =
dd->port->allocated[n/2] >> (32*(n&1));
else
group_allocated = dd->port->allocated[n];
- size += sprintf(&buf[size], "0x%08x\n",
- group_allocated);
+ size += sprintf(&buf[size], "%08X ", group_allocated);
}
+ size += sprintf(&buf[size], "]\n");
- size += sprintf(&buf[size], "Completed:\n");
-
- for (n = 0; n < dd->slot_groups; n++)
- size += sprintf(&buf[size], "0x%08x\n",
- readl(dd->port->completed[n]));
+ size += sprintf(&buf[size], "Commands in Q: [ 0x");
- size += sprintf(&buf[size], "PORT IRQ STAT : 0x%08x\n",
- readl(dd->port->mmio + PORT_IRQ_STAT));
- size += sprintf(&buf[size], "HOST IRQ STAT : 0x%08x\n",
- readl(dd->mmio + HOST_IRQ_STAT));
+ for (n = dd->slot_groups-1; n >= 0; n--) {
+ if (sizeof(long) > sizeof(u32))
+ group_allocated =
+ dd->port->cmds_to_issue[n/2] >> (32*(n&1));
+ else
+ group_allocated = dd->port->cmds_to_issue[n];
+ size += sprintf(&buf[size], "%08X ", group_allocated);
+ }
+ size += sprintf(&buf[size], "]\n");
return size;
}
@@ -2592,8 +2637,24 @@ static ssize_t mtip_hw_show_status(struct device *dev,
return size;
}
+static ssize_t mtip_hw_show_flags(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct driver_data *dd = dev_to_disk(dev)->private_data;
+ int size = 0;
+
+ size += sprintf(&buf[size], "Flag in port struct : [ %08lX ]\n",
+ dd->port->flags);
+ size += sprintf(&buf[size], "Flag in dd struct : [ %08lX ]\n",
+ dd->dd_flag);
+
+ return size;
+}
+
static DEVICE_ATTR(registers, S_IRUGO, mtip_hw_show_registers, NULL);
static DEVICE_ATTR(status, S_IRUGO, mtip_hw_show_status, NULL);
+static DEVICE_ATTR(flags, S_IRUGO, mtip_hw_show_flags, NULL);
/*
* Create the sysfs related attributes.
@@ -2616,6 +2677,9 @@ static int mtip_hw_sysfs_init(struct driver_data *dd, struct kobject *kobj)
if (sysfs_create_file(kobj, &dev_attr_status.attr))
dev_warn(&dd->pdev->dev,
"Error creating 'status' sysfs entry\n");
+ if (sysfs_create_file(kobj, &dev_attr_flags.attr))
+ dev_warn(&dd->pdev->dev,
+ "Error creating 'flags' sysfs entry\n");
return 0;
}
@@ -2636,6 +2700,7 @@ static int mtip_hw_sysfs_exit(struct driver_data *dd, struct kobject *kobj)
sysfs_remove_file(kobj, &dev_attr_registers.attr);
sysfs_remove_file(kobj, &dev_attr_status.attr);
+ sysfs_remove_file(kobj, &dev_attr_flags.attr);
return 0;
}
@@ -3634,7 +3699,10 @@ skip_create_disk:
set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
blk_queue_physical_block_size(dd->queue, 4096);
+ blk_queue_max_hw_sectors(dd->queue, 0xffff);
+ blk_queue_max_segment_size(dd->queue, 0x400000);
blk_queue_io_min(dd->queue, 4096);
+
/*
* write back cache is not supported in the device. FUA depends on
* write back cache support, hence setting flush support to zero.
diff --git a/drivers/block/mtip32xx/mtip32xx.h b/drivers/block/mtip32xx/mtip32xx.h
index 4ef58336310a..b2c88da26b2a 100644
--- a/drivers/block/mtip32xx/mtip32xx.h
+++ b/drivers/block/mtip32xx/mtip32xx.h
@@ -113,33 +113,35 @@
#define __force_bit2int (unsigned int __force)
-/* below are bit numbers in 'flags' defined in mtip_port */
-#define MTIP_PF_IC_ACTIVE_BIT 0 /* pio/ioctl */
-#define MTIP_PF_EH_ACTIVE_BIT 1 /* error handling */
-#define MTIP_PF_SE_ACTIVE_BIT 2 /* secure erase */
-#define MTIP_PF_DM_ACTIVE_BIT 3 /* download microcde */
-#define MTIP_PF_PAUSE_IO ((1 << MTIP_PF_IC_ACTIVE_BIT) | \
+enum {
+ /* below are bit numbers in 'flags' defined in mtip_port */
+ MTIP_PF_IC_ACTIVE_BIT = 0, /* pio/ioctl */
+ MTIP_PF_EH_ACTIVE_BIT = 1, /* error handling */
+ MTIP_PF_SE_ACTIVE_BIT = 2, /* secure erase */
+ MTIP_PF_DM_ACTIVE_BIT = 3, /* download microcde */
+ MTIP_PF_PAUSE_IO = ((1 << MTIP_PF_IC_ACTIVE_BIT) | \
(1 << MTIP_PF_EH_ACTIVE_BIT) | \
(1 << MTIP_PF_SE_ACTIVE_BIT) | \
- (1 << MTIP_PF_DM_ACTIVE_BIT))
-
-#define MTIP_PF_SVC_THD_ACTIVE_BIT 4
-#define MTIP_PF_ISSUE_CMDS_BIT 5
-#define MTIP_PF_REBUILD_BIT 6
-#define MTIP_PF_SVC_THD_STOP_BIT 8
-
-/* below are bit numbers in 'dd_flag' defined in driver_data */
-#define MTIP_DDF_REMOVE_PENDING_BIT 1
-#define MTIP_DDF_OVER_TEMP_BIT 2
-#define MTIP_DDF_WRITE_PROTECT_BIT 3
-#define MTIP_DDF_STOP_IO ((1 << MTIP_DDF_REMOVE_PENDING_BIT) | \
+ (1 << MTIP_PF_DM_ACTIVE_BIT)),
+
+ MTIP_PF_SVC_THD_ACTIVE_BIT = 4,
+ MTIP_PF_ISSUE_CMDS_BIT = 5,
+ MTIP_PF_REBUILD_BIT = 6,
+ MTIP_PF_SVC_THD_STOP_BIT = 8,
+
+ /* below are bit numbers in 'dd_flag' defined in driver_data */
+ MTIP_DDF_REMOVE_PENDING_BIT = 1,
+ MTIP_DDF_OVER_TEMP_BIT = 2,
+ MTIP_DDF_WRITE_PROTECT_BIT = 3,
+ MTIP_DDF_STOP_IO = ((1 << MTIP_DDF_REMOVE_PENDING_BIT) | \
(1 << MTIP_DDF_OVER_TEMP_BIT) | \
- (1 << MTIP_DDF_WRITE_PROTECT_BIT))
+ (1 << MTIP_DDF_WRITE_PROTECT_BIT)),
-#define MTIP_DDF_CLEANUP_BIT 5
-#define MTIP_DDF_RESUME_BIT 6
-#define MTIP_DDF_INIT_DONE_BIT 7
-#define MTIP_DDF_REBUILD_FAILED_BIT 8
+ MTIP_DDF_CLEANUP_BIT = 5,
+ MTIP_DDF_RESUME_BIT = 6,
+ MTIP_DDF_INIT_DONE_BIT = 7,
+ MTIP_DDF_REBUILD_FAILED_BIT = 8,
+};
__packed struct smart_attr{
u8 attr_id;
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index 013c7a549fb6..65665c9c42c6 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -141,7 +141,7 @@ struct rbd_request {
struct rbd_snap {
struct device dev;
const char *name;
- size_t size;
+ u64 size;
struct list_head node;
u64 id;
};
@@ -175,8 +175,7 @@ struct rbd_device {
/* protects updating the header */
struct rw_semaphore header_rwsem;
char snap_name[RBD_MAX_SNAP_NAME_LEN];
- u32 cur_snap; /* index+1 of current snapshot within snap context
- 0 - for the head */
+ u64 snap_id; /* current snapshot id */
int read_only;
struct list_head node;
@@ -241,7 +240,7 @@ static void rbd_put_dev(struct rbd_device *rbd_dev)
put_device(&rbd_dev->dev);
}
-static int __rbd_update_snaps(struct rbd_device *rbd_dev);
+static int __rbd_refresh_header(struct rbd_device *rbd_dev);
static int rbd_open(struct block_device *bdev, fmode_t mode)
{
@@ -450,7 +449,9 @@ static void rbd_client_release(struct kref *kref)
struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref);
dout("rbd_release_client %p\n", rbdc);
+ spin_lock(&rbd_client_list_lock);
list_del(&rbdc->node);
+ spin_unlock(&rbd_client_list_lock);
ceph_destroy_client(rbdc->client);
kfree(rbdc->rbd_opts);
@@ -463,9 +464,7 @@ static void rbd_client_release(struct kref *kref)
*/
static void rbd_put_client(struct rbd_device *rbd_dev)
{
- spin_lock(&rbd_client_list_lock);
kref_put(&rbd_dev->rbd_client->kref, rbd_client_release);
- spin_unlock(&rbd_client_list_lock);
rbd_dev->rbd_client = NULL;
}
@@ -487,16 +486,18 @@ static void rbd_coll_release(struct kref *kref)
*/
static int rbd_header_from_disk(struct rbd_image_header *header,
struct rbd_image_header_ondisk *ondisk,
- int allocated_snaps,
+ u32 allocated_snaps,
gfp_t gfp_flags)
{
- int i;
- u32 snap_count;
+ u32 i, snap_count;
if (memcmp(ondisk, RBD_HEADER_TEXT, sizeof(RBD_HEADER_TEXT)))
return -ENXIO;
snap_count = le32_to_cpu(ondisk->snap_count);
+ if (snap_count > (UINT_MAX - sizeof(struct ceph_snap_context))
+ / sizeof (*ondisk))
+ return -EINVAL;
header->snapc = kmalloc(sizeof(struct ceph_snap_context) +
snap_count * sizeof (*ondisk),
gfp_flags);
@@ -506,11 +507,11 @@ static int rbd_header_from_disk(struct rbd_image_header *header,
header->snap_names_len = le64_to_cpu(ondisk->snap_names_len);
if (snap_count) {
header->snap_names = kmalloc(header->snap_names_len,
- GFP_KERNEL);
+ gfp_flags);
if (!header->snap_names)
goto err_snapc;
header->snap_sizes = kmalloc(snap_count * sizeof(u64),
- GFP_KERNEL);
+ gfp_flags);
if (!header->snap_sizes)
goto err_names;
} else {
@@ -552,21 +553,6 @@ err_snapc:
return -ENOMEM;
}
-static int snap_index(struct rbd_image_header *header, int snap_num)
-{
- return header->total_snaps - snap_num;
-}
-
-static u64 cur_snap_id(struct rbd_device *rbd_dev)
-{
- struct rbd_image_header *header = &rbd_dev->header;
-
- if (!rbd_dev->cur_snap)
- return 0;
-
- return header->snapc->snaps[snap_index(header, rbd_dev->cur_snap)];
-}
-
static int snap_by_name(struct rbd_image_header *header, const char *snap_name,
u64 *seq, u64 *size)
{
@@ -605,7 +591,7 @@ static int rbd_header_set_snap(struct rbd_device *dev, u64 *size)
snapc->seq = header->snap_seq;
else
snapc->seq = 0;
- dev->cur_snap = 0;
+ dev->snap_id = CEPH_NOSNAP;
dev->read_only = 0;
if (size)
*size = header->image_size;
@@ -613,8 +599,7 @@ static int rbd_header_set_snap(struct rbd_device *dev, u64 *size)
ret = snap_by_name(header, dev->snap_name, &snapc->seq, size);
if (ret < 0)
goto done;
-
- dev->cur_snap = header->total_snaps - ret;
+ dev->snap_id = snapc->seq;
dev->read_only = 1;
}
@@ -935,7 +920,6 @@ static int rbd_do_request(struct request *rq,
layout->fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
layout->fl_stripe_count = cpu_to_le32(1);
layout->fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
- layout->fl_pg_preferred = cpu_to_le32(-1);
layout->fl_pg_pool = cpu_to_le32(dev->poolid);
ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
req, ops);
@@ -1168,7 +1152,7 @@ static int rbd_req_read(struct request *rq,
int coll_index)
{
return rbd_do_op(rq, rbd_dev, NULL,
- (snapid ? snapid : CEPH_NOSNAP),
+ snapid,
CEPH_OSD_OP_READ,
CEPH_OSD_FLAG_READ,
2,
@@ -1187,7 +1171,7 @@ static int rbd_req_sync_read(struct rbd_device *dev,
u64 *ver)
{
return rbd_req_sync_op(dev, NULL,
- (snapid ? snapid : CEPH_NOSNAP),
+ snapid,
CEPH_OSD_OP_READ,
CEPH_OSD_FLAG_READ,
NULL,
@@ -1238,7 +1222,7 @@ static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
dout("rbd_watch_cb %s notify_id=%lld opcode=%d\n", dev->obj_md_name,
notify_id, (int)opcode);
mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
- rc = __rbd_update_snaps(dev);
+ rc = __rbd_refresh_header(dev);
mutex_unlock(&ctl_mutex);
if (rc)
pr_warning(RBD_DRV_NAME "%d got notification but failed to "
@@ -1521,7 +1505,7 @@ static void rbd_rq_fn(struct request_queue *q)
coll, cur_seg);
else
rbd_req_read(rq, rbd_dev,
- cur_snap_id(rbd_dev),
+ rbd_dev->snap_id,
ofs,
op_size, bio,
coll, cur_seg);
@@ -1592,7 +1576,7 @@ static int rbd_read_header(struct rbd_device *rbd_dev,
{
ssize_t rc;
struct rbd_image_header_ondisk *dh;
- int snap_count = 0;
+ u32 snap_count = 0;
u64 ver;
size_t len;
@@ -1656,7 +1640,7 @@ static int rbd_header_add_snap(struct rbd_device *dev,
struct ceph_mon_client *monc;
/* we should create a snapshot only if we're pointing at the head */
- if (dev->cur_snap)
+ if (dev->snap_id != CEPH_NOSNAP)
return -EINVAL;
monc = &dev->rbd_client->client->monc;
@@ -1683,7 +1667,9 @@ static int rbd_header_add_snap(struct rbd_device *dev,
if (ret < 0)
return ret;
- dev->header.snapc->seq = new_snapid;
+ down_write(&dev->header_rwsem);
+ dev->header.snapc->seq = new_snapid;
+ up_write(&dev->header_rwsem);
return 0;
bad:
@@ -1703,7 +1689,7 @@ static void __rbd_remove_all_snaps(struct rbd_device *rbd_dev)
/*
* only read the first part of the ondisk header, without the snaps info
*/
-static int __rbd_update_snaps(struct rbd_device *rbd_dev)
+static int __rbd_refresh_header(struct rbd_device *rbd_dev)
{
int ret;
struct rbd_image_header h;
@@ -1890,7 +1876,7 @@ static ssize_t rbd_image_refresh(struct device *dev,
mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
- rc = __rbd_update_snaps(rbd_dev);
+ rc = __rbd_refresh_header(rbd_dev);
if (rc < 0)
ret = rc;
@@ -1949,7 +1935,7 @@ static ssize_t rbd_snap_size_show(struct device *dev,
{
struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
- return sprintf(buf, "%zd\n", snap->size);
+ return sprintf(buf, "%llu\n", (unsigned long long)snap->size);
}
static ssize_t rbd_snap_id_show(struct device *dev,
@@ -1958,7 +1944,7 @@ static ssize_t rbd_snap_id_show(struct device *dev,
{
struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
- return sprintf(buf, "%llu\n", (unsigned long long) snap->id);
+ return sprintf(buf, "%llu\n", (unsigned long long)snap->id);
}
static DEVICE_ATTR(snap_size, S_IRUGO, rbd_snap_size_show, NULL);
@@ -2173,7 +2159,7 @@ static int rbd_init_watch_dev(struct rbd_device *rbd_dev)
rbd_dev->header.obj_version);
if (ret == -ERANGE) {
mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
- rc = __rbd_update_snaps(rbd_dev);
+ rc = __rbd_refresh_header(rbd_dev);
mutex_unlock(&ctl_mutex);
if (rc < 0)
return rc;
@@ -2558,7 +2544,7 @@ static ssize_t rbd_snap_add(struct device *dev,
if (ret < 0)
goto err_unlock;
- ret = __rbd_update_snaps(rbd_dev);
+ ret = __rbd_refresh_header(rbd_dev);
if (ret < 0)
goto err_unlock;
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index 0d39f2f4294a..693187df7601 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -29,9 +29,6 @@ struct virtio_blk
/* The disk structure for the kernel. */
struct gendisk *disk;
- /* Request tracking. */
- struct list_head reqs;
-
mempool_t *pool;
/* Process context for config space updates */
@@ -55,7 +52,6 @@ struct virtio_blk
struct virtblk_req
{
- struct list_head list;
struct request *req;
struct virtio_blk_outhdr out_hdr;
struct virtio_scsi_inhdr in_hdr;
@@ -99,7 +95,6 @@ static void blk_done(struct virtqueue *vq)
}
__blk_end_request_all(vbr->req, error);
- list_del(&vbr->list);
mempool_free(vbr, vblk->pool);
}
/* In case queue is stopped waiting for more buffers. */
@@ -184,7 +179,6 @@ static bool do_req(struct request_queue *q, struct virtio_blk *vblk,
return false;
}
- list_add_tail(&vbr->list, &vblk->reqs);
return true;
}
@@ -437,7 +431,6 @@ static int __devinit virtblk_probe(struct virtio_device *vdev)
goto out_free_index;
}
- INIT_LIST_HEAD(&vblk->reqs);
spin_lock_init(&vblk->lock);
vblk->vdev = vdev;
vblk->sg_elems = sg_elems;
@@ -583,21 +576,29 @@ static void __devexit virtblk_remove(struct virtio_device *vdev)
{
struct virtio_blk *vblk = vdev->priv;
int index = vblk->index;
+ struct virtblk_req *vbr;
+ unsigned long flags;
/* Prevent config work handler from accessing the device. */
mutex_lock(&vblk->config_lock);
vblk->config_enable = false;
mutex_unlock(&vblk->config_lock);
- /* Nothing should be pending. */
- BUG_ON(!list_empty(&vblk->reqs));
-
/* Stop all the virtqueues. */
vdev->config->reset(vdev);
flush_work(&vblk->config_work);
del_gendisk(vblk->disk);
+
+ /* Abort requests dispatched to driver. */
+ spin_lock_irqsave(&vblk->lock, flags);
+ while ((vbr = virtqueue_detach_unused_buf(vblk->vq))) {
+ __blk_end_request_all(vbr->req, -EIO);
+ mempool_free(vbr, vblk->pool);
+ }
+ spin_unlock_irqrestore(&vblk->lock, flags);
+
blk_cleanup_queue(vblk->disk->queue);
put_disk(vblk->disk);
mempool_destroy(vblk->pool);
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 4e86393a09cf..60eed4bdd2e4 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -526,6 +526,14 @@ static int xen_translate_vdev(int vdevice, int *minor, unsigned int *offset)
return 0;
}
+static char *encode_disk_name(char *ptr, unsigned int n)
+{
+ if (n >= 26)
+ ptr = encode_disk_name(ptr, n / 26 - 1);
+ *ptr = 'a' + n % 26;
+ return ptr + 1;
+}
+
static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
struct blkfront_info *info,
u16 vdisk_info, u16 sector_size)
@@ -536,6 +544,7 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
unsigned int offset;
int minor;
int nr_parts;
+ char *ptr;
BUG_ON(info->gd != NULL);
BUG_ON(info->rq != NULL);
@@ -560,7 +569,11 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
"emulated IDE disks,\n\t choose an xvd device name"
"from xvde on\n", info->vdevice);
}
- err = -ENODEV;
+ if (minor >> MINORBITS) {
+ pr_warn("blkfront: %#x's minor (%#x) out of range; ignoring\n",
+ info->vdevice, minor);
+ return -ENODEV;
+ }
if ((minor % nr_parts) == 0)
nr_minors = nr_parts;
@@ -574,23 +587,14 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
if (gd == NULL)
goto release;
- if (nr_minors > 1) {
- if (offset < 26)
- sprintf(gd->disk_name, "%s%c", DEV_NAME, 'a' + offset);
- else
- sprintf(gd->disk_name, "%s%c%c", DEV_NAME,
- 'a' + ((offset / 26)-1), 'a' + (offset % 26));
- } else {
- if (offset < 26)
- sprintf(gd->disk_name, "%s%c%d", DEV_NAME,
- 'a' + offset,
- minor & (nr_parts - 1));
- else
- sprintf(gd->disk_name, "%s%c%c%d", DEV_NAME,
- 'a' + ((offset / 26) - 1),
- 'a' + (offset % 26),
- minor & (nr_parts - 1));
- }
+ strcpy(gd->disk_name, DEV_NAME);
+ ptr = encode_disk_name(gd->disk_name + sizeof(DEV_NAME) - 1, offset);
+ BUG_ON(ptr >= gd->disk_name + DISK_NAME_LEN);
+ if (nr_minors > 1)
+ *ptr = 0;
+ else
+ snprintf(ptr, gd->disk_name + DISK_NAME_LEN - ptr,
+ "%d", minor & (nr_parts - 1));
gd->major = XENVBD_MAJOR;
gd->first_minor = minor;
@@ -1496,7 +1500,9 @@ module_init(xlblk_init);
static void __exit xlblk_exit(void)
{
- return xenbus_unregister_driver(&blkfront_driver);
+ xenbus_unregister_driver(&blkfront_driver);
+ unregister_blkdev(XENVBD_MAJOR, DEV_NAME);
+ kfree(minors);
}
module_exit(xlblk_exit);
generated by cgit v1.2.3 (git 2.39.0) at 2024-12-28 19:03:36 +0300