linux-kernel - Re: [PATCH 2/2] drivers: core: Remove glue dirs from sysfs earlier

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CA+55aFxqUyACDPBL4trTfGyJudg8P7YGazVGUW7-UaNFJiwZRQ@mail.gmail.com>
Date:   Sat, 30 Jun 2018 20:57:05 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Benjamin Herrenschmidt <benh@...nel.crashing.org>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Eric W. Biederman" <ebiederm@...ssion.com>,
        Joel Stanley <joel@....id.au>
Subject: Re: [PATCH 2/2] drivers: core: Remove glue dirs from sysfs earlier

On Sat, Jun 30, 2018 at 8:42 PM Benjamin Herrenschmidt
<benh@...nel.crashing.org> wrote:
>
> But what if something *else* still holds a reference to the kobject ?
> It could be anything really... t

But that's fine. Then the object will continue to exist, and the sysfs
file will continue to exist, and you won't get a new glue directory.

In fact, what you describe is a problem with *your* patch, exactly
because you introduce another counter that is *not* the reference
count, and now you have the problem with "old directory kobject is
still live, but I removed the sysfs part, and now I'm creating a new
object with the same name".

Hmm?

> It is and there's a WARN_ON about it inside kobject_get(). I don't
> think anybody argues against that, you are absolutely right.

No.  That the zero kobject_get() will not result in a warning. It just
does a kref_get(), no warnings anywhere.

Yes, there is a kobject_get() warning, but that's about an
_uninitialized_ kobject, not a already released one! You will get no
warning that I can see from the "oh, you just got a stale one".

           Linus