drbd: Bitmap IO functions can now return prematurely if the disk breaks
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index a2c337b3..e5e756d 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -886,7 +886,7 @@
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;
@@ -897,6 +897,7 @@
{
struct bm_aio_ctx *ctx = container_of(kref, struct bm_aio_ctx, kref);
+ put_ldev(ctx->mdev);
kfree(ctx);
}
@@ -945,7 +946,8 @@
bio_put(bio);
if (atomic_dec_and_test(&ctx->in_flight)) {
- complete(&ctx->done);
+ ctx->done = 1;
+ wake_up(&mdev->misc_wait);
kref_put(&ctx->kref, &bm_aio_ctx_destroy);
}
}
@@ -1034,12 +1036,18 @@
*ctx = (struct bm_aio_ctx) {
.mdev = mdev,
.in_flight = ATOMIC_INIT(1),
- .done = COMPLETION_INITIALIZER(ctx->done),
+ .done = 0,
.flags = lazy_writeout_upper_idx ? BM_AIO_COPY_PAGES : 0,
.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));
@@ -1073,11 +1081,16 @@
/*
* 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
+ * 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);
+ wait_until_done_or_disk_failure(mdev, &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",
@@ -1089,6 +1102,9 @@
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);
@@ -1103,7 +1119,6 @@
ppsize(ppb, now << (BM_BLOCK_SHIFT-10)), now);
kref_put(&ctx->kref, &bm_aio_ctx_destroy);
-
return err;
}
@@ -1167,14 +1182,20 @@
*ctx = (struct bm_aio_ctx) {
.mdev = mdev,
.in_flight = ATOMIC_INIT(1),
- .done = COMPLETION_INITIALIZER(ctx->done),
+ .done = 0,
.flags = BM_AIO_COPY_PAGES,
.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 drbd_bm_write_page()\n");
+ kfree(ctx);
+ return -ENODEV;
+ }
+
bm_page_io_async(ctx, idx, WRITE_SYNC);
- wait_for_completion(&ctx->done);
+ wait_until_done_or_disk_failure(mdev, &ctx->done);
if (ctx->error)
drbd_chk_io_error(mdev, 1, true);
@@ -1182,9 +1203,8 @@
* gone in a moment as well. */
mdev->bm_writ_cnt++;
- err = ctx->error;
+ err = atomic_read(&ctx->in_flight) ? -EIO : ctx->error;
kref_put(&ctx->kref, &bm_aio_ctx_destroy);
-
return err;
}