Skip to contentSkip to conversation

f2fs: prevent compressed file from being disabled after releasing cblocks

PATCH In review Daeho Jeong 3 participants 10 messages Aug 28, 2020 – Aug 31, 2020 F2FS file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 f2fs: prevent compressed file from being disabled after releasing cblocks

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

---
 fs/f2fs/f2fs.h | 2 ++
 1 file changed, 2 insertions(+)
fs/f2fs/f2fs.h +2 -0
index 02811ce15059..14d30740ba03 100644
@@ -3936,6 +3936,8 @@ static inline u64 f2fs_disable_compressed_file(struct inode *inode)
 	if (!f2fs_compressed_file(inode))
 		return 0;
 	if (S_ISREG(inode->i_mode)) {
+		if (IS_IMMUTABLE(inode))
+			return 1;
 		if (get_dirty_pages(inode))
 			return 1;
 		if (fi->i_compr_blocks)
Signature · 1 line
2.28.0.402.g5ffc5be6b7-goog
maintainer
#1 lore raw
Quoted 17 lines · Daeho Jeong

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

---
 fs/f2fs/f2fs.h | 2 ++
 1 file changed, 2 insertions(+)
fs/f2fs/f2fs.h +0 -0
index 02811ce15059..14d30740ba03 100644

fs/f2fs/f2fs.h static inline u64 f2fs_disable_compressed_file(struct inode *inode) · line 3940

Quoted 5 lines
fs/f2fs/f2fs.h +2 -0
@@ -3936,6 +3936,8 @@ static inline u64 f2fs_disable_compressed_file(struct inode *inode)
 	if (!f2fs_compressed_file(inode))
 		return 0;
 	if (S_ISREG(inode->i_mode)) {
+		if (IS_IMMUTABLE(inode))
+			return 1;

It looks most of callers are from ioctl, should we add immutable check in f2fs ioctl interfaces if necessary? or I missed existed check.

Thanks,

Quoted 4 lines
fs/f2fs/f2fs.h +0 -0
 		if (get_dirty_pages(inode))
 			return 1;
 		if (fi->i_compr_blocks)
#2 lore raw

Do you have any reason not to put this check here?
If we do this check outside of here, we definitely make a mistake sooner or later.

2020년 8월 30일 (일) 오후 12:24, Chao Yu <[email protected]>님이 작성:

Quoted 28 lines · Daeho Jeong
Quoted 17 lines

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

---
 fs/f2fs/f2fs.h | 2 ++
 1 file changed, 2 insertions(+)
fs/f2fs/f2fs.h +0 -0
index 02811ce15059..14d30740ba03 100644

fs/f2fs/f2fs.h static inline u64 f2fs_disable_compressed_file(struct inode *inode) · line 3940

Quoted 5 lines
fs/f2fs/f2fs.h +2 -0
@@ -3936,6 +3936,8 @@ static inline u64 f2fs_disable_compressed_file(struct inode *inode)
      if (!f2fs_compressed_file(inode))
              return 0;
      if (S_ISREG(inode->i_mode)) {
+             if (IS_IMMUTABLE(inode))
+                     return 1;

It looks most of callers are from ioctl, should we add immutable check in f2fs ioctl interfaces if necessary? or I missed existed check.

Quoted 6 lines

Thanks,

Quoted 4 lines
fs/f2fs/f2fs.h +0 -0
              if (get_dirty_pages(inode))
                      return 1;
              if (fi->i_compr_blocks)

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

#3 lore raw
Quoted 1 line · Daeho Jeong

Do you have any reason not to put this check here?

No, the place is okay to me. :)

Quoted 2 lines

If we do this check outside of here, we definitely make a mistake sooner or later.

I just want to see whether we can cover all cases in where we missed to add immutable check condition if necessary.

Thanks,

Quoted 43 lines

2020년 8월 30일 (일) 오후 12:24, Chao Yu <[email protected]>님이 작성:

Quoted 28 lines · Daeho Jeong
Quoted 17 lines

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

---
  fs/f2fs/f2fs.h | 2 ++
  1 file changed, 2 insertions(+)
fs/f2fs/f2fs.h +0 -0
index 02811ce15059..14d30740ba03 100644

fs/f2fs/f2fs.h static inline u64 f2fs_disable_compressed_file(struct inode *inode) · line 3940

Quoted 5 lines
fs/f2fs/f2fs.h +2 -0
@@ -3936,6 +3936,8 @@ static inline u64 f2fs_disable_compressed_file(struct inode *inode)
       if (!f2fs_compressed_file(inode))
               return 0;
       if (S_ISREG(inode->i_mode)) {
+             if (IS_IMMUTABLE(inode))
+                     return 1;

It looks most of callers are from ioctl, should we add immutable check in f2fs ioctl interfaces if necessary? or I missed existed check.

Quoted 6 lines

Thanks,

Quoted 4 lines
fs/f2fs/f2fs.h +0 -0
               if (get_dirty_pages(inode))
                       return 1;
               if (fi->i_compr_blocks)

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

#4 lore raw

Sorry, I didn't get your point.

So, do you think this patch is ok? And we need to consider that we need more immutable checks for other cases?
Or you want to remove this immutable check from here and add the check to each ioctl functions?

2020년 8월 31일 (월) 오전 10:24, Chao Yu <[email protected]>님이 작성:

Quoted 57 lines · Daeho Jeong
Quoted 1 line

Do you have any reason not to put this check here?

No, the place is okay to me. :)

Quoted 2 lines

If we do this check outside of here, we definitely make a mistake sooner or later.

I just want to see whether we can cover all cases in where we missed to add immutable check condition if necessary.

Thanks,

Quoted 43 lines

2020년 8월 30일 (일) 오후 12:24, Chao Yu <[email protected]>님이 작성:

Quoted 28 lines · Daeho Jeong
Quoted 17 lines

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

---
  fs/f2fs/f2fs.h | 2 ++
  1 file changed, 2 insertions(+)
fs/f2fs/f2fs.h +0 -0
index 02811ce15059..14d30740ba03 100644

fs/f2fs/f2fs.h static inline u64 f2fs_disable_compressed_file(struct inode *inode) · line 3940

Quoted 5 lines
fs/f2fs/f2fs.h +2 -0
@@ -3936,6 +3936,8 @@ static inline u64 f2fs_disable_compressed_file(struct inode *inode)
       if (!f2fs_compressed_file(inode))
               return 0;
       if (S_ISREG(inode->i_mode)) {
+             if (IS_IMMUTABLE(inode))
+                     return 1;

It looks most of callers are from ioctl, should we add immutable check in f2fs ioctl interfaces if necessary? or I missed existed check.

Quoted 6 lines

Thanks,

Quoted 4 lines
fs/f2fs/f2fs.h +0 -0
               if (get_dirty_pages(inode))
                       return 1;
               if (fi->i_compr_blocks)

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

#5 lore raw
Quoted 4 lines · Daeho Jeong

Sorry, I didn't get your point.

So, do you think this patch is ok? And we need to consider that we need more immutable checks for other cases?

Yes, this patch looks good to me.

But, IMO, we should discuss about whether we need to add more immutable checks for other ioctl cases.

- open(O_RDWR)
- ioctl(FS_IOC_SETFLAGS, F2FS_COMPR_FL)
- write()
- ioctl(RELEASE_COMPRESS_BLOCKS) -- inode is immutable now
- ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL) -- Should we allow to update immutable inode? as we know, normally, immutable inode should deny open(O_WRONLY or O_RDWR) and later update.

Thanks,

Quoted 62 lines

Or you want to remove this immutable check from here and add the check to each ioctl functions? >
2020년 8월 31일 (월) 오전 10:24, Chao Yu <[email protected]>님이 작성:

Quoted 6 lines · Daeho Jeong
Quoted 1 line

Do you have any reason not to put this check here?

No, the place is okay to me. :)

Quoted 51 lines
Quoted 2 lines

If we do this check outside of here, we definitely make a mistake sooner or later.

I just want to see whether we can cover all cases in where we missed to add immutable check condition if necessary.

Thanks,

Quoted 43 lines

2020년 8월 30일 (일) 오후 12:24, Chao Yu <[email protected]>님이 작성:

Quoted 28 lines · Daeho Jeong
Quoted 17 lines

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

---
   fs/f2fs/f2fs.h | 2 ++
   1 file changed, 2 insertions(+)
fs/f2fs/f2fs.h +0 -0
index 02811ce15059..14d30740ba03 100644

fs/f2fs/f2fs.h static inline u64 f2fs_disable_compressed_file(struct inode *inode) · line 3940

Quoted 5 lines
fs/f2fs/f2fs.h +2 -0
@@ -3936,6 +3936,8 @@ static inline u64 f2fs_disable_compressed_file(struct inode *inode)
        if (!f2fs_compressed_file(inode))
                return 0;
        if (S_ISREG(inode->i_mode)) {
+             if (IS_IMMUTABLE(inode))
+                     return 1;

It looks most of callers are from ioctl, should we add immutable check in f2fs ioctl interfaces if necessary? or I missed existed check.

Quoted 6 lines

Thanks,

Quoted 4 lines
fs/f2fs/f2fs.h +0 -0
                if (get_dirty_pages(inode))
                        return 1;
                if (fi->i_compr_blocks)

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

.

#6 lore raw
Quoted 7 lines

- open(O_RDWR)
- ioctl(FS_IOC_SETFLAGS, F2FS_COMPR_FL)
- write()
- ioctl(RELEASE_COMPRESS_BLOCKS) -- inode is immutable now
- ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL) -- Should we allow to update immutable inode? as we know, normally, immutable inode should deny open(O_WRONLY or O_RDWR) and later update.

For this case, with this patch we'll return -EINVAL for ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL).
I thought RESERVE_COMPRESS_BLOCKS ioctl is always required to get the file to normal mode after RELEASE_COMPRESS_BLOCKS is called.

#7 lore raw
Quoted 8 lines · Daeho Jeong
Quoted 7 lines

- open(O_RDWR)
- ioctl(FS_IOC_SETFLAGS, F2FS_COMPR_FL)
- write()
- ioctl(RELEASE_COMPRESS_BLOCKS) -- inode is immutable now
- ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL) -- Should we allow to update immutable inode? as we know, normally, immutable inode should deny open(O_WRONLY or O_RDWR) and later update.

Quoted 4 lines

For this case, with this patch we'll return -EINVAL for ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL).
I thought RESERVE_COMPRESS_BLOCKS ioctl is always required to get the file to normal mode after RELEASE_COMPRESS_BLOCKS is called.

That's an example, after compressed block release, shouldn't we disallow other ioctl interface which updates immutable inode?

Quoted 2 lines

.

#8 lore raw

I think this patch is enough for now.
If you have anything else we should take care, please let me know~

2020년 8월 31일 (월) 오후 3:08, Chao Yu <[email protected]>님이 작성:

Quoted 20 lines · Daeho Jeong
Quoted 8 lines
Quoted 7 lines

- open(O_RDWR)
- ioctl(FS_IOC_SETFLAGS, F2FS_COMPR_FL)
- write()
- ioctl(RELEASE_COMPRESS_BLOCKS) -- inode is immutable now
- ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL) -- Should we allow to update immutable inode? as we know, normally, immutable inode should deny open(O_WRONLY or O_RDWR) and later update.

Quoted 4 lines

For this case, with this patch we'll return -EINVAL for ioctl(FS_IOC_SETFLAGS, ~F2FS_COMPR_FL).
I thought RESERVE_COMPRESS_BLOCKS ioctl is always required to get the file to normal mode after RELEASE_COMPRESS_BLOCKS is called.

That's an example, after compressed block release, shouldn't we disallow other ioctl interface which updates immutable inode?

Quoted 2 lines

.

Reviewed-by on v1 1/1
#9 lore raw
In reply to Daeho Jeong
Quoted 8 lines · Daeho Jeong

From: Daeho Jeong <[email protected]>

After releasing cblocks, the compressed file can be accidentally disabled in compression mode, since it has zero cblocks. As we are using IMMUTABLE flag to present released cblocks state, we can add IMMUTABLE state check when considering the compressed file disabling.

Thanks,