On Mon, Apr 26, 2021 at 7:58 PM Darrick J. Wong <djwong@xxxxxxxxxx> wrote:
>
> Please pull this single patch to the iomap code for 5.13-rc1, which
> augments what gets logged when someone tries to swapon an unacceptable
> swap file. (Yes, this is a continuation of the swapfile drama from last
> season...)
Hmm. I've pulled this, but that "iomap_swapfile_fail()" thing seems a
bit silly to me.
We have '%pD' for printing a filename. It may not be perfect (by
default it only prints one component, you can do "%pD4" to show up to
four components), but it should "JustWork(tm)".
And if it doesn't, we should fix it.
So instead of having a kmalloc/kfree for the path buffer, I think you
should have been able to just do
pr_err("swapon: file %pD4 %s\n", isi->file, str);
and be done with it.
And no, we don't have a ton of %pD users, so if it's ugly or buggy
when the file is NULL, or has problems with more (of fewer) than four
path components, let's just fix that (added Jia He and Al Viro to
participants, they've been the two people doing %pd and %pD - for
'struct dentry *' and 'struct file *' respectively).
Linus
