[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160324160905.GS29764@suse.cz>
Date: Thu, 24 Mar 2016 17:09:06 +0100
From: David Sterba <dsterba@...e.cz>
To: Petr Tesarik <ptesarik@...e.com>
Cc: Flex Liu <fliu@...ell.com>, David Sterba <dsterba@...e.com>,
linux-btrfs@...r.kernel.org, Chris Mason <clm@...com>,
Josef Bacik <jbacik@...com>, linux-kernel@...r.kernel.org,
Flex Liu <fliu@...e.com>
Subject: Re: [PATCH 1/1] Btrfs: Code Cleanup
On Thu, Mar 24, 2016 at 04:08:18PM +0100, Petr Tesarik wrote:
> On Thu, 24 Mar 2016 16:03:07 +0100
> David Sterba <dsterba@...e.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.
That's acceptable, if there are going to be incremental improvements.
Powered by blists - more mailing lists