linux-kernel - Re: [PATCH 0/4] remove PDE

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]

Date:   Tue, 16 Nov 2021 16:26:12 +0800
From:   Muchun Song <songmuchun@...edance.com>
To:     Andrew Morton <akpm@...ux-foundation.org>
Cc:     Alexey Dobriyan <adobriyan@...il.com>, gladkov.alexey@...il.com,
        LKML <linux-kernel@...r.kernel.org>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>
Subject: Re: [PATCH 0/4] remove PDE_DATA()

On Tue, Nov 16, 2021 at 1:09 PM Andrew Morton <akpm@...ux-foundation.org> wrote:
>
> On Mon,  1 Nov 2021 17:35:14 +0800 Muchun Song <songmuchun@...edance.com> wrote:
>
> > I found a bug [1] some days ago, which is because we want to use
> > inode->i_private to pass user private data. However, this is wrong
> > on proc fs. We provide a specific function PDE_DATA() to get user
> > private data. Actually, we can hide this detail by storing
> > PDE()->data into inode->i_private and removing PDE_DATA() completely.
> > The user could use inode->i_private to get user private data just
> > like debugfs does. This series is trying to remove PDE_DATA().
>
> Why can't we do
>
> /*
>  * comment goes here
>  */
> static inline void *PDE_DATA(struct inode *inode)
> {
>         return inode->i_private;
> }
>
> to abstract things a bit and to reduce the patch size?
>
> otoh, that upper-case thing needs to go, so the patch size remains the
> same anyway.
>
> And perhaps we should have a short-term
>
> #define PDE_DATA(i) pde_data(i)

Right. This way is the easiest way to reduce the patch size.
Actually, I want to make all PDE_DATA() go away, which
makes this patch series go big.

>
> because new instances are sure to turn up during the development cycle.
>
> But I can handle that by staging the patch series after linux-next and
> reminding myself to grep for new PDE_DATA instances prior to
> upstreaming.

I'd be happy if you could replace PDE_DATA() with inode->i_private.
In this case, should I still introduce pde_data() and perform the above
things to make this series smaller?

Thanks.