Commit 074f4afe authored by Lars Ellenberg's avatar Lars Ellenberg Committed by Jens Axboe

drbd: fix a race between start_resync and send_and_submit

In the drbd make request function, specifically in
drbd_send_and_submit(), we decide whether we want to send the actual
write request, or only a "set this block out of sync" information.

We do so based on the current connection state, while holding the req_lock.
The connection state is not supposed to change while holding the req_lock.

But in drbd_start_resync, we did change that state anyways,
while only holding the global_state_lock, which is enough to change
sync-after dependencies (paused vs active resync), but
not good enough to change the connection state.

Fix: in drbd_start_resync, first grab the req_lock to serialize with
drbd_send_and_submit(), before grabbing the global_state_lock
to be able to evaluate the sync-after dependencies.
Signed-off-by: default avatarPhilipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: default avatarLars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: default avatarJens Axboe <axboe@fb.com>
parent 20c68fde
......@@ -1686,11 +1686,15 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
}
clear_bit(B_RS_H_DONE, &device->flags);
write_lock_irq(&global_state_lock);
/* req_lock: serialize with drbd_send_and_submit() and others
* global_state_lock: for stable sync-after dependencies */
spin_lock_irq(&device->resource->req_lock);
write_lock(&global_state_lock);
/* Did some connection breakage or IO error race with us? */
if (device->state.conn < C_CONNECTED
|| !get_ldev_if_state(device, D_NEGOTIATING)) {
write_unlock_irq(&global_state_lock);
write_unlock(&global_state_lock);
spin_unlock_irq(&device->resource->req_lock);
mutex_unlock(device->state_mutex);
return;
}
......@@ -1730,7 +1734,8 @@ void drbd_start_resync(struct drbd_device *device, enum drbd_conns side)
}
_drbd_pause_after(device);
}
write_unlock_irq(&global_state_lock);
write_unlock(&global_state_lock);
spin_unlock_irq(&device->resource->req_lock);
if (r == SS_SUCCESS) {
/* reset rs_last_bcast when a resync or verify is started,
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment