linux-kernel - Re: [PATCH 0/2] btrfs: Prevent open-coded arithmetic on kmalloc

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: <20250923061144.GS5333@twin.jikos.cz>
Date: Tue, 23 Sep 2025 08:11:44 +0200
From: David Sterba <dsterba@...e.cz>
To: Miquel Sabaté Solà <mssola@...ola.com>
Cc: David Sterba <dsterba@...e.cz>, linux-btrfs@...r.kernel.org, clm@...com,
	dsterba@...e.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/2] btrfs: Prevent open-coded arithmetic on kmalloc

On Mon, Sep 22, 2025 at 02:51:15PM +0200, Miquel Sabaté Solà wrote:
> > On Fri, Sep 19, 2025 at 04:58:14PM +0200, Miquel Sabaté Solà wrote:
> >> The second patch is a small cleanup after fixing up my first patch, in
> >> which I realized that the __free(kfree) attribute would come in handy in a
> >> couple of particularly large functions with multiple exit points. This
> >> second patch is probably more of a cosmetic thing, and it's not an
> >> exhaustive exercise by any means. All of this to say that even if I feel
> >> like it should be included, I don't mind if it has to be dropped.
> >
> > Yes there are many candidates for the __free() cleanup annotation and
> > we'll want to fix them all systematically. We already have the automatic
> > cleaning for struct btrfs_path (BTRFS_PATH_AUTO_FREE). For the
> > kfree/kvfree I'd like to something similar:
> >
> > #define AUTO_KFREE(name)       *name __free(kfree) = NULL
> > #define AUTO_KVFREE(name)      *name __free(kvfree) = NULL
> >
> > This wraps the name and initializes it to NULL so it's not accidentally
> > forgotten.
> 
> Makes sense! I can take a look at this if nobody else is working on it,
> even if I think it should go into a new patch series.

Thanks, it's yours. Yes this should be in a separate patchset.

> Hence, if it sounds good to you, we can merge this patch as it is right
> now, and in parallel I work on this proposed AUTO_KFREE and AUTO_KVFREE
> macros in a new patch series (which will take more time to prepare).

I'd rather see all the changes done the same way so it's not __free and
then converted to AUTO_KFREE. Also the development branch is frozen
before 6.18 pull request so all that will be in the 6.19 cycle anyway.