linux-kernel - Re: [f2fs-dev] [PATCH v4 2/2] f2fs: fix compat F2FS_IOC_{MOVE, GARBAGE_COLLECT}

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:   Sat, 7 Nov 2020 09:16:35 -0800
From:   Eric Biggers <ebiggers@...nel.org>
To:     Chao Yu <chao@...nel.org>
Cc:     Chao Yu <yuchao0@...wei.com>, jaegeuk@...nel.org,
        linux-kernel@...r.kernel.org,
        linux-f2fs-devel@...ts.sourceforge.net
Subject: Re: [f2fs-dev] [PATCH v4 2/2] f2fs: fix compat F2FS_IOC_{MOVE,
 GARBAGE_COLLECT}_RANGE

On Sat, Nov 07, 2020 at 05:25:23PM +0800, Chao Yu wrote:
> On 2020/11/7 2:03, Eric Biggers wrote:
> > On Fri, Nov 06, 2020 at 02:53:31PM +0800, Chao Yu wrote:
> > > +#if defined(__KERNEL__)
> > > +struct compat_f2fs_gc_range {
> > > +	u32 sync;
> > > +	compat_u64 start;
> > > +	compat_u64 len;
> > > +};
> > 
> > There's no need to use '#if defined(__KERNEL__)' in kernel source files.
> > 
> > Likewise for compat_f2fs_move_range.
> 
> Correct.
> 
> > 
> > > +static int f2fs_compat_ioc_gc_range(struct file *file, unsigned long arg)
> > > +{
> > > +	struct f2fs_sb_info *sbi = F2FS_I_SB(file_inode(file));
> > > +	struct compat_f2fs_gc_range __user *urange;
> > > +	struct f2fs_gc_range range;
> > > +	int err;
> > > +
> > > +	if (unlikely(f2fs_cp_error(sbi)))
> > > +		return -EIO;
> > > +	if (!f2fs_is_checkpoint_ready(sbi))
> > > +		return -ENOSPC;
> > 
> > I still don't understand why this checkpoint-related stuff is getting added
> > here, and only to the compat versions of the ioctls.  It wasn't in the original
> > version.  If they are needed then they should be added to __f2fs_ioc_gc_range()
> > and __f2fs_ioc_move_range() (preferably by a separate patch) so that they are
> 
> If so, cp-related stuff will be checked redundantly in both f2fs_ioctl() and
> __f2fs_ioc_xxx() function for native GC_RANGE and MOVE_RANGE ioctls, it's
> not needed.
> 

Oh I see, the cp-related checks are at the beginning of f2fs_ioctl() too.

In that case a much better approach would be to add __f2fs_ioctl() which is
called by f2fs_ioctl() and f2fs_compat_ioctl(), and have f2fs_ioctl() and
f2fs_compat_ioctl() do the cp-related checks but not __f2fs_ioctl().

I feel that's still not entirely correct, because ENOTTY should take precedence
over EIO or ENOSPC from the cp-related stuff.  But at least it would be
consistent between the native and compat ioctls, and the cp-related checks
wouldn't have to be duplicated in random ioctls...

- Eric