Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH] f2fs: fix potential .flags overflow on 32bit architecture
To:     =?UTF-8?Q?Ond=c5=99ej_Jirman?= <megi@xff.cz>,
        Chao Yu <chao@kernel.org>, <jaegeuk@kernel.org>,
        <linux-f2fs-devel@lists.sourceforge.net>,
        <linux-kernel@vger.kernel.org>
References: <20200322101327.5979-1-chao@kernel.org>
 <20200322121434.i2jea6o5tzanip7z@core.my.home>
 <47c71fe9-e168-8080-d0ed-2cfaa9a77e5e@kernel.org>
 <20200322153014.czpca7ozhfy4ctdh@core.my.home>
From:   Chao Yu <yuchao0@huawei.com>
Message-ID: <70f992f5-54f7-bd09-2b27-f418da51341e@huawei.com>
Date:   Mon, 23 Mar 2020 09:15:39 +0800
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101
 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <20200322153014.czpca7ozhfy4ctdh@core.my.home>
Content-Type: text/plain; charset="utf-8"
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 2020/3/22 23:30, Ondřej Jirman wrote:
> On Sun, Mar 22, 2020 at 09:18:56PM +0800, Chao Yu wrote:
>> Hi,
>>
>> On 2020-3-22 20:14, Ondřej Jirman wrote:
>>> Hello,
>>>
>>> On Sun, Mar 22, 2020 at 06:13:27PM +0800, Chao Yu wrote:
>>>> From: Chao Yu <yuchao0@huawei.com>
>>>>
>>>> f2fs_inode_info.flags is unsigned long variable, it has 32 bits
>>>> in 32bit architecture, since we introduced FI_MMAP_FILE flag
>>>> when we support data compression, we may access memory cross
>>>> the border of .flags field, corrupting .i_sem field, result in
>>>> below deadlock.
>>>>
>>>> To fix this issue, let's introduce .extra_flags to grab extra
>>>> space to store those new flags.
>>>>
>>>> Call Trace:
>>>>  __schedule+0x8d0/0x13fc
>>>>  ? mark_held_locks+0xac/0x100
>>>>  schedule+0xcc/0x260
>>>>  rwsem_down_write_slowpath+0x3ab/0x65d
>>>>  down_write+0xc7/0xe0
>>>>  f2fs_drop_nlink+0x3d/0x600 [f2fs]
>>>>  f2fs_delete_inline_entry+0x300/0x440 [f2fs]
>>>>  f2fs_delete_entry+0x3a1/0x7f0 [f2fs]
>>>>  f2fs_unlink+0x500/0x790 [f2fs]
>>>>  vfs_unlink+0x211/0x490
>>>>  do_unlinkat+0x483/0x520
>>>>  sys_unlink+0x4a/0x70
>>>>  do_fast_syscall_32+0x12b/0x683
>>>>  entry_SYSENTER_32+0xaa/0x102
>>>>
>>>> Fixes: 4c8ff7095bef ("f2fs: support data compression")
>>>> Signed-off-by: Chao Yu <yuchao0@huawei.com>
>>>> ---
>>>>  fs/f2fs/f2fs.h  | 26 ++++++++++++++++++++------
>>>>  fs/f2fs/inode.c |  1 +
>>>>  2 files changed, 21 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
>>>> index fcafa68212eb..fcd22df2e9ca 100644
>>>> --- a/fs/f2fs/f2fs.h
>>>> +++ b/fs/f2fs/f2fs.h
>>>> @@ -695,6 +695,7 @@ struct f2fs_inode_info {
>>>>
>>>>  	/* Use below internally in f2fs*/
>>>>  	unsigned long flags;		/* use to pass per-file flags */
>>>> +	unsigned long extra_flags;	/* extra flags */
>>>>  	struct rw_semaphore i_sem;	/* protect fi info */
>>>>  	atomic_t dirty_pages;		/* # of dirty pages */
>>>>  	f2fs_hash_t chash;		/* hash value of given file name */
>>>> @@ -2569,7 +2570,7 @@ enum {
>>>>  };
>>>>
>>>>  static inline void __mark_inode_dirty_flag(struct inode *inode,
>>>> -						int flag, bool set)
>>>> +					unsigned long long flag, bool set)
>>>>  {
>>>>  	switch (flag) {
>>>>  	case FI_INLINE_XATTR:
>>>> @@ -2588,20 +2589,33 @@ static inline void __mark_inode_dirty_flag(struct inode *inode,
>>>>
>>>>  static inline void set_inode_flag(struct inode *inode, int flag)
>>>>  {
>>>> -	if (!test_bit(flag, &F2FS_I(inode)->flags))
>>>> -		set_bit(flag, &F2FS_I(inode)->flags);
>>>> +	if ((1 << flag) <= sizeof(unsigned long)) {
>>>
>>> ^ this is wrong. Maybe you meant flag <= BITS_PER_LONG
>>
>> Oh, my bad, I meant that, thanks for pointing out this. :)
>>
>>>
>>> And ditto for the same checks below. Maybe you can make flags an array of
>>> BIT_WORD(max_flag_value) + 1 and skip the branches altogether?
> 
> I've found maybe a clearer macro for this: 
> 
>   https://elixir.bootlin.com/linux/latest/source/include/linux/bitops.h#L15
> 
>   BITS_TO_LONGS(nr)

More clean, thanks, will send v3 soon.

Thanks,

> 
> But it takes the number of bits in the bitmap, which would be 
> "max_flag_value + 1".
> 
> regards,
> 	o.
> 
>> That will be better, let me revise this patch.
>>
>> Thanks,
>>
>>>
>>> thank you and regards,
>>> 	o.
>>>
>>>> +		if (!test_bit(flag, &F2FS_I(inode)->flags))
>>>> +			set_bit(flag, &F2FS_I(inode)->flags);
>>>> +	} else {
>>>> +		if (!test_bit(flag - 32, &F2FS_I(inode)->extra_flags))
>>>> +			set_bit(flag - 32, &F2FS_I(inode)->extra_flags);
>>>> +	}
>>>>  	__mark_inode_dirty_flag(inode, flag, true);
>>>>  }
>>>>
>>>>  static inline int is_inode_flag_set(struct inode *inode, int flag)
>>>>  {
>>>> -	return test_bit(flag, &F2FS_I(inode)->flags);
>>>> +	if ((1 << flag) <= sizeof(unsigned long))
>>>> +		return test_bit(flag, &F2FS_I(inode)->flags);
>>>> +	else
>>>> +		return test_bit(flag - 32, &F2FS_I(inode)->extra_flags);
>>>>  }
>>>>
>>>>  static inline void clear_inode_flag(struct inode *inode, int flag)
>>>>  {
>>>> -	if (test_bit(flag, &F2FS_I(inode)->flags))
>>>> -		clear_bit(flag, &F2FS_I(inode)->flags);
>>>> +	if ((1 << flag) <= sizeof(unsigned long)) {
>>>> +		if (test_bit(flag, &F2FS_I(inode)->flags))
>>>> +			clear_bit(flag, &F2FS_I(inode)->flags);
>>>> +	} else {
>>>> +		if (test_bit(flag - 32, &F2FS_I(inode)->extra_flags))
>>>> +			clear_bit(flag - 32, &F2FS_I(inode)->extra_flags);
>>>> +	}
>>>>  	__mark_inode_dirty_flag(inode, flag, false);
>>>>  }
>>>>
>>>> diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
>>>> index 44e08bf2e2b4..ca924d7e0e30 100644
>>>> --- a/fs/f2fs/inode.c
>>>> +++ b/fs/f2fs/inode.c
>>>> @@ -363,6 +363,7 @@ static int do_read_inode(struct inode *inode)
>>>>  	if (S_ISREG(inode->i_mode))
>>>>  		fi->i_flags &= ~F2FS_PROJINHERIT_FL;
>>>>  	fi->flags = 0;
>>>> +	fi->extra_flags = 0;
>>>>  	fi->i_advise = ri->i_advise;
>>>>  	fi->i_pino = le32_to_cpu(ri->i_pino);
>>>>  	fi->i_dir_level = ri->i_dir_level;
>>>> --
>>>> 2.22.0
>>>>
> .
>