Re: [PATCH] ext4: Log inode exhaustion to dmesg

[Theodore Ts'o <tytso@mit.edu>]

Hi,

The Signed-off-by needs to contain your real name (sorry, no
pseudonyms or anonymous contributions.)  That's because it has a
formal legal meaning.  See Section 11: "Sign your work - the
Developer’s Certificate of Origin" of the Submitting Patches
docuementation:

https://www.kernel.org/doc/html/latest/process/submitting-patches.html

(Also, you should use ext4_warning() and not a bare printk).

						- Ted

On Mon, Oct 23, 2017 at 05:32:47PM +0530, Team Athena wrote:
> Make a log in dmesg when file creation fails due to no free inodes.
> The error code for both "out of disk space" and "out of inode" is the same.
> This is misleading to the user. Logging the exact reason helps to find and
> correct the issue from the users' side.
> 
> Fix bug 197335 - https://bugzilla.kernel.org/show_bug.cgi?id=197335
> 
> Signed-off-by: Team Athena <teamathena.nitc@gmail.com>
> ---
>  fs/ext4/namei.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
> index c1cf020d..c3990d2d 100644
> --- a/fs/ext4/namei.c
> +++ b/fs/ext4/namei.c
> @@ -2463,6 +2463,8 @@ static int ext4_create(struct inode *dir, struct dentry *dentry, umode_t mode,
>  		ext4_journal_stop(handle);
>  	if (err == -ENOSPC && ext4_should_retry_alloc(dir->i_sb, &retries))
>  		goto retry;
> +	else if (err == -ENOSPC && printk_ratelimited())
> +		printk(pr_warning "ext4: No space on disk, inode usage full");
>  	return err;
>  }
>  
> -- 
> 2.11.0
>