On Thu, 24 Mar 2016 16:03:07 +0100
David Sterba <dste...@suse.cz> wrote:
> On Sun, Mar 20, 2016 at 03:11:11PM +0800, Flex Liu wrote:
>[...]
> > --- a/fs/btrfs/volumes.c
> > +++ b/fs/btrfs/volumes.c
> > @@ -2325,7 +2325,10 @@ int btrfs_init_new_device(struct btrfs_root *root,
> > char *device_path)
> > if (seeding_dev) {
> > sb->s_flags &= ~MS_RDONLY;
> > ret = btrfs_prepare_sprout(root);
> > - BUG_ON(ret); /* -ENOMEM */
> > + if (ret) {
> > + btrfs_abort_transaction(trans, root, ret);
>
> The transaction abort seems a bit heavy as it will take down the whole
> filesystem. It's called from the device add ioctl, this is a restartable
> operation.
>
> Unfortunatelly btrfs_prepare_sprout is called after the transaction
> start so btrfs_abort_transaction must be called. To avoid it, the code
> would need to be reorganized, so the memory allocations happen in
> advance.
On the other hand, an abort is still better than a BUG_ON(), and it may
be easier to make incremental improvements.
Just my 2 cents (I haven't tried it),
Petr T
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
