linux-kernel - Re: [PATCH] vfs: fix statfs64() returning impossible EOVERFLOW for 64-bit f

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:   Mon, 6 Aug 2018 11:45:29 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Al Viro <viro@...iv.linux.org.uk>
Cc:     Sergey Klyaus <sergey.m.klyaus@...il.com>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Li Wang <liwang@...hat.com>,
        Andreas Dilger <adilger@...ger.ca>,
        Andi Kleen <andi.kleen@...el.com>
Subject: Re: [PATCH] vfs: fix statfs64() returning impossible EOVERFLOW for
 64-bit f_files

On Mon, Aug 6, 2018 at 10:06 AM Al Viro <viro@...iv.linux.org.uk> wrote:
>
> That leaves us with f_bsize and f_frsize (the latter defaulting to the former).
> Hugetlbfs can put greater than 4Gb values in there, for really huge pages.
> And that's the only thing worth checking in there.
>
> So the whole put_compat_statfs64() thing should be

Ack, I'm ok with this simplification.

> I'm somewhat tempted to get rid of those 'long' in struct kstatfs,

I'm ok with this one too.

> with
>
> #define STATFS_COPYOUT(type)                                            \
> static int put##type(struct kstatfs *st, struct type __user *p)         \

No. Don't do this.

I'm ok with the #define to avoid duplication, but don't bother with
the FIT_IN() after you've above successfully argued that it's
pointless for anything but f_bsize/frsize.

So if you do the macro to generate the different copyout versions,
just use your simplified code for that macro instead. With FIT_IN()
just for f_bsize/frsize.

                    Linus