Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20201029041538.4165209-1-daeho43@gmail.com> <20201029041538.4165209-2-daeho43@gmail.com>
 <92f4da8e-27a1-7577-84f9-39038eaa88cb@huawei.com> <20201029155453.GB849@sol.localdomain>
In-Reply-To: <20201029155453.GB849@sol.localdomain>
From:   Daeho Jeong <daeho43@gmail.com>
Date:   Fri, 30 Oct 2020 10:46:16 +0900
Message-ID: <CACOAw_xc5KYhVgp-b8NTs8vapWxiLF3m-qUMO2qgjp4gQQb2Pw@mail.gmail.com>
Subject: Re: [f2fs-dev] [PATCH v5 2/2] f2fs: add F2FS_IOC_SET_COMPRESS_OPTION ioctl
To:     Eric Biggers <ebiggers@kernel.org>
Cc:     Chao Yu <yuchao0@huawei.com>, linux-kernel@vger.kernel.org,
        linux-f2fs-devel@lists.sourceforge.net, kernel-team@android.com,
        Daeho Jeong <daehojeong@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Chao, got it on f2fs_is_compress_algorithm_valid().
I also agree on Eric's opinion on that error value.

2020=EB=85=84 10=EC=9B=94 30=EC=9D=BC (=EA=B8=88) =EC=98=A4=EC=A0=84 12:54,=
 Eric Biggers <ebiggers@kernel.org>=EB=8B=98=EC=9D=B4 =EC=9E=91=EC=84=B1:
>
> On Thu, Oct 29, 2020 at 03:29:17PM +0800, Chao Yu wrote:
> > > +static int f2fs_ioc_set_compress_option(struct file *filp, unsigned =
long arg)
> > > +{
> > > +   struct inode *inode =3D file_inode(filp);
> > > +   struct f2fs_sb_info *sbi =3D F2FS_I_SB(inode);
> > > +   struct f2fs_comp_option option;
> > > +   int ret =3D 0;
> > > +
> > > +   if (!f2fs_sb_has_compression(sbi))
> > > +           return -EOPNOTSUPP;
> > > +
> > > +   if (!(filp->f_mode & FMODE_WRITE))
> > > +           return -EBADF;
> > > +
> > > +   if (copy_from_user(&option, (struct f2fs_comp_option __user *)arg=
,
> > > +                           sizeof(option)))
> > > +           return -EFAULT;
> > > +
> > > +   if (!f2fs_compressed_file(inode) ||
> > > +                   option.log_cluster_size < MIN_COMPRESS_LOG_SIZE |=
|
> > > +                   option.log_cluster_size > MAX_COMPRESS_LOG_SIZE |=
|
> > > +                   option.algorithm >=3D COMPRESS_MAX)
> > > +           return -EINVAL;
> > > +
> > > +   file_start_write(filp);
> > > +   inode_lock(inode);
> > > +
> > > +   if (f2fs_is_mmap_file(inode) || get_dirty_pages(inode)) {
> > > +           ret =3D -EBUSY;
> > > +           goto out;
> > > +   }
> > > +
> > > +   if (inode->i_size !=3D 0) {
> > > +           ret =3D -EFBIG;
> > > +           goto out;
> > > +   }
> >
> > Hmm...
> >
> > Shouldn't it be:
> >
> > if (algorithm >=3D COMPRESS_MAX) {
> >       ret =3D -ENOPKG;
> >       goto out;
> > }
> >
> > if (!f2fs_cops[algorithm])
> >       f2fs_warn(...);
>
> Note that my intent with recommending ENOPKG was for it to be returned in=
 the
> !f2fs_cops[algorithm] case, similar to how opening an encrypted file when=
 the
> encryption algorithm is recognized but not supported by the kernel return=
s
> ENOPKG.  For a truly unrecognized algorithm (algorithm >=3D COMPRESS_MAX)=
, EINVAL
> would probably be more appropriate.  So if !f2fs_cops[algorithm] is now a=
llowed,
> then ENOPKG should no longer be among the error codes this ioctl returns.
>
> - Eric