linux-ext4 - Re: [PATCH] ext3: Avoid creating new file in append-only dir when open(2) return error

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CA+dCu8_cHh7EeSnkO-KH7LM4aOqt0noHtO6ZyU4YS=Ff+BVChg@mail.gmail.com>
Date:	Wed, 2 Nov 2011 10:05:35 +0800
From:	Eryu Guan <guaneryu@...il.com>
To:	Jan Kara <jack@...e.cz>
Cc:	linux-ext4@...r.kernel.org
Subject: Re: [PATCH] ext3: Avoid creating new file in append-only dir when
 open(2) return error

On Wed, Nov 2, 2011 at 7:10 AM, Jan Kara <jack@...e.cz> wrote:
> On Sat 29-10-11 02:03:07, Eryu Guan wrote:
>> Newly created file on ext3 inherits inode flags from parent directory,
>> so new inode created in append-only directory has S_APPEND flag set,
>> may_open() called by do_last() checks that flag then returns -EPERM,
>> but at that time the new inode is already created.
>>
>> This can be reproduced by:
>>       # mkdir -p /mnt/ext3/append-only
>>       # chattr +a /mnt/ext3/append-only
>>       # ./opentest /mnt/ext3/append-only/newtestfile
>>       # ls -l /mnt/ext3/append-only/newtestfile
>>
>> opentest will return 'Operation not permitted', but the ls shows that
>> newtestfile is already created.
>>
>>       # cat opentest.c
>>       #include <stdio.h>
>>       #include <sys/types.h>
>>       #include <fcntl.h>
>>       #include <sys/stat.h>
>>
>>       int main(int argc, char *argv[])
>>       {
>>               int fd;
>>               fd = open(argv[1], O_RDWR|O_CREAT, 0666);
>>               if (fd == -1)
>>                       perror("open failed");
>>               return 0;
>>       }
>>
>> To avoid this, check EXT3_APPEND_FL flag first in ext3_create before
>> really allocating new inode.
>  Yes, it is nicer to not create any file when open(2) fails in the end.
> BTW, how have you spotted this? I've taken your ext2 and ext3 patches into
> my tree.

I found this by xfstests 079, since it's no longer xfs specific and
fails on extN.
After applying my patch, 079 still fails on extN, but not leave
unwanted file there.

As Ted said, the other way to fix this is to remove EXTx_APPEND_FL from
EXTx_FL_INHERITED, so new inode will not inherit the append-only flag. This fix
will also make 079 pass on extN.

Though I prefer the later way, I still sent out my patch for review,
because it's
not a behavior-changer.

Thanks.
Eryu Guan
>
>                                                                Honza
>>
>> Cc: Jan Kara <jack@...e.cz>
>> Signed-off-by: Eryu Guan <guaneryu@...il.com>
>> ---
>>  fs/ext3/namei.c |   10 ++++++++++
>>  1 files changed, 10 insertions(+), 0 deletions(-)
>>
>> diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
>> index 0629e09..323cf2f 100644
>> --- a/fs/ext3/namei.c
>> +++ b/fs/ext3/namei.c
>> @@ -36,6 +36,7 @@
>>  #include <linux/quotaops.h>
>>  #include <linux/buffer_head.h>
>>  #include <linux/bio.h>
>> +#include <linux/namei.h>
>>  #include <trace/events/ext3.h>
>>
>>  #include "namei.h"
>> @@ -1704,6 +1705,15 @@ static int ext3_create (struct inode * dir, struct dentry * dentry, int mode,
>>       handle_t *handle;
>>       struct inode * inode;
>>       int err, retries = 0;
>> +     int open_flag = nd->intent.open.file->f_flags;
>> +
>> +     if ((EXT3_I(dir)->i_flags & EXT3_FL_INHERITED) & EXT3_APPEND_FL) {
>> +             if ((open_flag & O_ACCMODE) != O_RDONLY &&
>> +                 !(open_flag & O_APPEND))
>> +                     return -EPERM;
>> +             if (open_flag & O_TRUNC)
>> +                     return -EPERM;
>> +     }
>>
>>       dquot_initialize(dir);
>>
>> --
>> 1.7.7.1
>>
> --
> Jan Kara <jack@...e.cz>
> SUSE Labs, CR
>
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html