Re: [PATCH v2] btrfs: fix infinite directory reads

[Eugeniu Rosca <erosca@de.adit-jv.com>]

Hi Filipe and Qu,

On Thu, Jan 25, 2024 at 10:02:01AM +0000, Filipe Manana wrote:
> On Thu, Jan 25, 2024 at 9:51 AM Qu Wenruo <wqu@suse.com> wrote:
> >
> > From: Filipe Manana <fdmanana@suse.com>
> >
> > [ Upstream commit 9b378f6ad48cfa195ed868db9123c09ee7ec5ea2 ]
> >
> > The readdir implementation currently processes always up to the last index
> > it finds. This however can result in an infinite loop if the directory has

[..]

> Thanks for the backport, and running the corresponding test case from
> fstests to verify it's working.
> 
> However when backporting a commit, one should also check if there are
> fixes for that commit, as they
> often introduce regressions or have some other bug - 

+1. Good to see this best practice applied here.

> and that's the
> case here. We also need to backport
> the following 3 commits:
> 
> https:// git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=357950361cbc6d54fb68ed878265c647384684ae
> https:// git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=e60aa5da14d01fed8411202dbe4adf6c44bd2a57
> https:// git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8e7f82deb0c0386a03b62e30082574347f8b57d5

Good catch. I get the same list thanks to the reference of the culprit:

$ git log --oneline --grep 9b378f6ad linux/master
8e7f82deb0c038 btrfs: fix race between reading a directory and adding entries to it
e60aa5da14d01f btrfs: refresh dir last index during a rewinddir(3) call
357950361cbc6d btrfs: set last dir index to the current last index when opening dir

> One regression, the one regarding rewinddir(3), even has a test case
> in fstests too (generic/471) and would have been caught
> when running the "dir" group tests in fstests:
> 
> https:// git.kernel.org/pub/scm/fs/xfs/xfstests-dev.git/commit/?h=for-next&id=68b958f5dc4ab13cfd86f7fb82621f9f022b7626
> 
> I'll work on making backports of those 3 other patches on top of your
> backport, and then send all of them in a series,
> including your patch, to make it easier to follow and apply all at once.

Thanks for your support. Looking forward.

BR, Eugeniu