Re: ext4 regression in v5.9-rc2 from e7bfb5c9bb3d on ro fs with overlapped bitmaps
From: Josh Triplett
Date: Thu Oct 08 2020 - 18:28:15 EST
On Thu, Oct 08, 2020 at 01:25:57PM -0600, Andreas Dilger wrote:
> On Oct 8, 2020, at 1:12 PM, Josh Triplett <josh@xxxxxxxxxxxxxxxx> wrote:
> > On Wed, Oct 07, 2020 at 08:57:12PM -0600, Andreas Dilger wrote:
> >> I *do* think that inline_data is an under-appreciated feature that I
> >> would be happy to see some improvements with. I don't think that small
> >> files are a niche use case, and if we can clean up the inline_data code
> >> to work with 128-byte inodes I'm not against that, even though I'm not
> >> going to use that combination of features myself.
> >
> > I'd love to see that happen. At the time, it seemed like too large of a
> > change to block on, which is why I ended up deciding to switch to
> > 256-byte inodes.
>
> Does that mean you are using inline_data with 256-byte inodes?
I am, yes, and it mostly works great. I've hit zero issues with it in
the filesystems I'm generating.
> That would also be good to know, since there haven't been any
> well-known users of this feature so far (AFAIK). Since you are using
> this in a read-only manner, you won't hit the one know issue when an
> inline_data inode is extended to use an external block that may
> temporarily leave the inode in an inconsistent state.
I've run into a few other issues with it in other tools, as well. mke2fs
with inline_data generates invalid files given xattrs:
https://lore.kernel.org/linux-ext4/20200926102512.GA11386@localhost/T/#u
And extlinux doesn't like inline_data at all:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=971002
I'll report any other issues I run into using inline_data. I agree that
it's deeply underappreciated.
- Josh