[PATCH] f2fs: Add offset check routine before punch_hole() in f2fs_fallocate()

[PATCH] f2fs: Add offset check routine before punch_hole() in f2fs_fallocate()

[Taehee Yoo]

In the punch_hole(), if offset bigger than inode size, it returns SUCCESS.
Then f2fs_fallocate() will update time and dirty mark.
In that case, inode has not been modified actually.
So I have added offset check routine that prevent to call the punch_hole().

Signed-off-by: Taehee Yoo <ap420073@gmail.com>
---
 fs/f2fs/file.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index a6f3f61..6470831 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -830,15 +830,19 @@ static long f2fs_fallocate(struct file *file, int mode,
 				loff_t offset, loff_t len)
 {
 	struct inode *inode = file_inode(file);
-	long ret;
+	long ret = 0;
 
 	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
 		return -EOPNOTSUPP;
 
 	mutex_lock(&inode->i_mutex);
 
-	if (mode & FALLOC_FL_PUNCH_HOLE)
+	if (mode & FALLOC_FL_PUNCH_HOLE) {
+		if (offset >= inode->i_size)
+			goto out;
+
 		ret = punch_hole(inode, offset, len);
+	}
 	else
 		ret = expand_inode_data(inode, offset, len, mode);
 
@@ -847,6 +851,7 @@ static long f2fs_fallocate(struct file *file, int mode,
 		mark_inode_dirty(inode);
 	}
 
+out:
 	mutex_unlock(&inode->i_mutex);
 
 	trace_f2fs_fallocate(inode, mode, offset, len, ret);
-- 
1.9.1


------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF

Re: [PATCH] f2fs: Add offset check routine before punch_hole() in f2fs_fallocate()

[Chao Yu]

Hi Taehee,

> -----Original Message-----
> From: Taehee Yoo [mailto:ap420073@gmail.com]
> Sent: Monday, April 20, 2015 4:46 PM
> To: jaegeuk@kernel.org; cm224.lee@samsung.com; linux-f2fs-devel@lists.sourceforge.net
> Subject: [f2fs-dev] [PATCH] f2fs: Add offset check routine before punch_hole() in
> f2fs_fallocate()
> 
> In the punch_hole(), if offset bigger than inode size, it returns SUCCESS.
> Then f2fs_fallocate() will update time and dirty mark.
> In that case, inode has not been modified actually.
> So I have added offset check routine that prevent to call the punch_hole().

Nice catch!

This patch makes size verification in punch_hole() redundant, how about removing
it?

Thanks,

> 
> Signed-off-by: Taehee Yoo <ap420073@gmail.com>
> ---
>  fs/f2fs/file.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> index a6f3f61..6470831 100644
> --- a/fs/f2fs/file.c
> +++ b/fs/f2fs/file.c
> @@ -830,15 +830,19 @@ static long f2fs_fallocate(struct file *file, int mode,
>  				loff_t offset, loff_t len)
>  {
>  	struct inode *inode = file_inode(file);
> -	long ret;
> +	long ret = 0;
> 
>  	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
>  		return -EOPNOTSUPP;
> 
>  	mutex_lock(&inode->i_mutex);
> 
> -	if (mode & FALLOC_FL_PUNCH_HOLE)
> +	if (mode & FALLOC_FL_PUNCH_HOLE) {
> +		if (offset >= inode->i_size)
> +			goto out;
> +
>  		ret = punch_hole(inode, offset, len);
> +	}
>  	else
>  		ret = expand_inode_data(inode, offset, len, mode);
> 
> @@ -847,6 +851,7 @@ static long f2fs_fallocate(struct file *file, int mode,
>  		mark_inode_dirty(inode);
>  	}
> 
> +out:
>  	mutex_unlock(&inode->i_mutex);
> 
>  	trace_f2fs_fallocate(inode, mode, offset, len, ret);
> --
> 1.9.1
> 



------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF

Re: [PATCH] f2fs: Add offset check routine before punch_hole() in f2fs_fallocate()

[Taehee Yoo]

2015-04-21 14:38 GMT+09:00 Chao Yu <chao2.yu@samsung.com>:
> Hi Taehee,
>
>> -----Original Message-----
>> From: Taehee Yoo [mailto:ap420073@gmail.com]
>> Sent: Monday, April 20, 2015 4:46 PM
>> To: jaegeuk@kernel.org; cm224.lee@samsung.com; linux-f2fs-devel@lists.sourceforge.net
>> Subject: [f2fs-dev] [PATCH] f2fs: Add offset check routine before punch_hole() in
>> f2fs_fallocate()
>>
>> In the punch_hole(), if offset bigger than inode size, it returns SUCCESS.
>> Then f2fs_fallocate() will update time and dirty mark.
>> In that case, inode has not been modified actually.
>> So I have added offset check routine that prevent to call the punch_hole().
>
> Nice catch!
>
> This patch makes size verification in punch_hole() redundant, how about removing
> it?
>
> Thanks,
>
>>
>> Signed-off-by: Taehee Yoo <ap420073@gmail.com>
>> ---
>>  fs/f2fs/file.c | 9 +++++++--
>>  1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
>> index a6f3f61..6470831 100644
>> --- a/fs/f2fs/file.c
>> +++ b/fs/f2fs/file.c
>> @@ -830,15 +830,19 @@ static long f2fs_fallocate(struct file *file, int mode,
>>                               loff_t offset, loff_t len)
>>  {
>>       struct inode *inode = file_inode(file);
>> -     long ret;
>> +     long ret = 0;
>>
>>       if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
>>               return -EOPNOTSUPP;
>>
>>       mutex_lock(&inode->i_mutex);
>>
>> -     if (mode & FALLOC_FL_PUNCH_HOLE)
>> +     if (mode & FALLOC_FL_PUNCH_HOLE) {
>> +             if (offset >= inode->i_size)
>> +                     goto out;
>> +
>>               ret = punch_hole(inode, offset, len);
>> +     }
>>       else
>>               ret = expand_inode_data(inode, offset, len, mode);
>>
>> @@ -847,6 +851,7 @@ static long f2fs_fallocate(struct file *file, int mode,
>>               mark_inode_dirty(inode);
>>       }
>>
>> +out:
>>       mutex_unlock(&inode->i_mutex);
>>
>>       trace_f2fs_fallocate(inode, mode, offset, len, ret);
>> --
>> 1.9.1
>>
>
>

I agree with your idea. and I will send V2 patch soon.

Thanks a lot for your review!

------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF