f2fs: fix data corruption issue with hardware encryption

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

---
 fs/f2fs/data.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

fs/f2fs/data.c +0 -0

index 9ef6f1f..7b2fef0 100644

fs/f2fs/data.c +7 -0

+		}
 		if (map->m_next_extent)
 			*map->m_next_extent = pgofs + map->m_len;
 		goto out;

@@ -1188,6 +1194,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 	goto next_dnode;
 
 sync_out:
+	/* for hardware encryption, but to avoid potential issue in future */
+	if (flag == F2FS_GET_BLOCK_DIO && map->m_flags & F2FS_MAP_MAPPED) {
+		blkaddr = map->m_pblk;
+		for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+			f2fs_wait_on_block_writeback(sbi, blkaddr);
+	}
 	if (flag == F2FS_GET_BLOCK_PRECACHE) {
 		if (map->m_flags & F2FS_MAP_MAPPED) {
 			unsigned int ofs = start_pgofs - map->m_lblk;

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Quoted 23 lines · Sahitya Tummala

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

---
 fs/f2fs/data.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

fs/f2fs/data.c int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, · line 1035

Quoted 13 lines

fs/f2fs/data.c +5 -0

index 9ef6f1f..7b2fef0 100644
@@ -1028,6 +1028,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		map->m_pblk = ei.blk + pgofs - ei.fofs;
 		map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
 		map->m_flags = F2FS_MAP_MAPPED;
+		/* for HW encryption, but to avoid potential issue in future */
+		if (flag == F2FS_GET_BLOCK_DIO) {
+			blkaddr = map->m_pblk;
+			for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+				f2fs_wait_on_block_writeback(sbi, blkaddr);

Do we need this? IIRC, DIO would give create=1.

Quoted 21 lines

fs/f2fs/data.c +7 -0

+		}
 		if (map->m_next_extent)
 			*map->m_next_extent = pgofs + map->m_len;
 		goto out;

@@ -1188,6 +1194,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 	goto next_dnode;
 
 sync_out:
+	/* for hardware encryption, but to avoid potential issue in future */
+	if (flag == F2FS_GET_BLOCK_DIO && map->m_flags & F2FS_MAP_MAPPED) {
+		blkaddr = map->m_pblk;
+		for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+			f2fs_wait_on_block_writeback(sbi, blkaddr);
+	}
 	if (flag == F2FS_GET_BLOCK_PRECACHE) {
 		if (map->m_flags & F2FS_MAP_MAPPED) {
 			unsigned int ofs = start_pgofs - map->m_lblk;

Signature · 2 lines

Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Quoted 24 lines · Jaegeuk Kim

Quoted 23 lines · Sahitya Tummala

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

---
 fs/f2fs/data.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

Quoted 15 lines

fs/f2fs/data.c int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, · line 1035

Quoted 13 lines

fs/f2fs/data.c +5 -0

index 9ef6f1f..7b2fef0 100644
@@ -1028,6 +1028,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		map->m_pblk = ei.blk + pgofs - ei.fofs;
 		map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
 		map->m_flags = F2FS_MAP_MAPPED;
+		/* for HW encryption, but to avoid potential issue in future */
+		if (flag == F2FS_GET_BLOCK_DIO) {
+			blkaddr = map->m_pblk;
+			for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+				f2fs_wait_on_block_writeback(sbi, blkaddr);

Do we need this? IIRC, DIO would give create=1.

Yes, we need it. When we are overwriting an existing file, DIO calls f2fs_map_blocks() with create=0. From the DIO code, I see that this happens because blockdev_direct_IO() passes this dio flag DIO_SKIP_HOLES. And then in get_more_blocks(), below code updates create=0, when we are overwriting an existing file.

                create = dio->op == REQ_OP_WRITE;
                if (dio->flags & DIO_SKIP_HOLES) {
                        if (fs_startblk <= ((i_size_read(dio->inode) - 1) >>
                                                        i_blkbits))
                                create = 0;
                }

                ret = (*sdio->get_block)(dio->inode, fs_startblk,
                                                map_bh, create);

From 216037253e6c21530ced716940fa01978d801f0d Mon Sep 17 00:00:00 2001

Quoted 54 lines · Sahitya Tummala

Quoted 24 lines · Jaegeuk Kim

Quoted 23 lines · Sahitya Tummala

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

---
 fs/f2fs/data.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

Quoted 15 lines

fs/f2fs/data.c int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, · line 1035

Quoted 13 lines

fs/f2fs/data.c +5 -0

index 9ef6f1f..7b2fef0 100644
@@ -1028,6 +1028,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		map->m_pblk = ei.blk + pgofs - ei.fofs;
 		map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
 		map->m_flags = F2FS_MAP_MAPPED;
+		/* for HW encryption, but to avoid potential issue in future */
+		if (flag == F2FS_GET_BLOCK_DIO) {
+			blkaddr = map->m_pblk;
+			for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+				f2fs_wait_on_block_writeback(sbi, blkaddr);

Do we need this? IIRC, DIO would give create=1.

Yes, we need it. When we are overwriting an existing file, DIO calls f2fs_map_blocks() with create=0. From the DIO code, I see that this happens because blockdev_direct_IO() passes this dio flag DIO_SKIP_HOLES. And then in get_more_blocks(), below code updates create=0, when we are overwriting an existing file.

                create = dio->op == REQ_OP_WRITE;
                if (dio->flags & DIO_SKIP_HOLES) {
                        if (fs_startblk <= ((i_size_read(dio->inode) - 1) >>
                                                        i_blkbits))
                                create = 0;
                }

Quoted 3 lines

                ret = (*sdio->get_block)(dio->inode, fs_startblk,
                                                map_bh, create);

Got it.
How about this?

From b78dd7b2e0317be18716b9496269e9792829f63e Mon Sep 17 00:00:00 2001

Quoted 59 lines · Jaegeuk Kim

Quoted 57 lines · Sahitya Tummala

Quoted 24 lines · Jaegeuk Kim

Quoted 23 lines · Sahitya Tummala

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

---
 fs/f2fs/data.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

Quoted 15 lines

fs/f2fs/data.c int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, · line 1035

Quoted 13 lines

fs/f2fs/data.c +5 -0

index 9ef6f1f..7b2fef0 100644
@@ -1028,6 +1028,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		map->m_pblk = ei.blk + pgofs - ei.fofs;
 		map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
 		map->m_flags = F2FS_MAP_MAPPED;
+		/* for HW encryption, but to avoid potential issue in future */
+		if (flag == F2FS_GET_BLOCK_DIO) {
+			blkaddr = map->m_pblk;
+			for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+				f2fs_wait_on_block_writeback(sbi, blkaddr);

Do we need this? IIRC, DIO would give create=1.

Yes, we need it. When we are overwriting an existing file, DIO calls f2fs_map_blocks() with create=0. From the DIO code, I see that this happens because blockdev_direct_IO() passes this dio flag DIO_SKIP_HOLES. And then in get_more_blocks(), below code updates create=0, when we are overwriting an existing file.

                create = dio->op == REQ_OP_WRITE;
                if (dio->flags & DIO_SKIP_HOLES) {
                        if (fs_startblk <= ((i_size_read(dio->inode) - 1) >>
                                                        i_blkbits))
                                create = 0;
                }

                ret = (*sdio->get_block)(dio->inode, fs_startblk,
                                                map_bh, create);

Quoted 3 lines

Got it.
How about this?

Sorry, this is v2.

From b78dd7b2e0317be18716b9496269e9792829f63e Mon Sep 17 00:00:00 2001
From: Sahitya Tummala <[email protected]>
Date: Wed, 10 Oct 2018 10:56:22 +0530
Subject: [PATCH] f2fs: fix data corruption issue with hardware encryption

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

Signed-off-by Jaegeuk Kim <[email protected]>

---
 fs/f2fs/data.c    | 11 +++++++++++
 fs/f2fs/f2fs.h    |  2 ++
 fs/f2fs/segment.c |  9 +++++++++
 3 files changed, 22 insertions(+)

fs/f2fs/data.c +11 -0

index be19257d9e36..8952f2d610a6 100644
@@ -1030,6 +1030,11 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		map->m_flags = F2FS_MAP_MAPPED;
 		if (map->m_next_extent)
 			*map->m_next_extent = pgofs + map->m_len;
+
+		/* for hardware encryption, but to avoid potential issue in future */
+		if (flag == F2FS_GET_BLOCK_DIO)
+			f2fs_wait_on_block_writeback_range(inode,
+						map->m_pblk, map->m_len);
 		goto out;
 	}
 

@@ -1188,6 +1193,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 	goto next_dnode;
 
 sync_out:
+
+	/* for hardware encryption, but to avoid potential issue in future */
+	if (flag == F2FS_GET_BLOCK_DIO && map->m_flags & F2FS_MAP_MAPPED)
+		f2fs_wait_on_block_writeback_range(inode,
+						map->m_pblk, map->m_len);
+
 	if (flag == F2FS_GET_BLOCK_PRECACHE) {
 		if (map->m_flags & F2FS_MAP_MAPPED) {
 			unsigned int ofs = start_pgofs - map->m_lblk;

fs/f2fs/f2fs.h +2 -0

index 7693b1a2072e..4a608a71c360 100644
@@ -2986,6 +2986,8 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
 void f2fs_wait_on_page_writeback(struct page *page,
 			enum page_type type, bool ordered);
 void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr);
+void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
+								block_t len);
 void f2fs_write_data_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
 void f2fs_write_node_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
 int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,

fs/f2fs/segment.c +9 -0

index 805c8310d7b0..f306d8946d99 100644
@@ -3301,6 +3301,15 @@ void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr)
 	}
 }
 
+void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
+								block_t len)
+{
+	block_t i;
+
+	for (i = 0; i < len; i++)
+		f2fs_wait_on_block_writeback(inode, blkaddr + i);
+}
+
 static int read_compacted_summaries(struct f2fs_sb_info *sbi)
 {
 	struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);

Signature · 1 line

2.19.0.605.g01d371f741-goog

Quoted 55 lines · Jaegeuk Kim

Quoted 54 lines · Sahitya Tummala

Quoted 24 lines · Jaegeuk Kim

Quoted 23 lines · Sahitya Tummala

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

---
 fs/f2fs/data.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

Quoted 15 lines

fs/f2fs/data.c int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, · line 1035

Quoted 13 lines

fs/f2fs/data.c +5 -0

index 9ef6f1f..7b2fef0 100644
@@ -1028,6 +1028,12 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
 		map->m_pblk = ei.blk + pgofs - ei.fofs;
 		map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
 		map->m_flags = F2FS_MAP_MAPPED;
+		/* for HW encryption, but to avoid potential issue in future */
+		if (flag == F2FS_GET_BLOCK_DIO) {
+			blkaddr = map->m_pblk;
+			for (; blkaddr < map->m_pblk + map->m_len; blkaddr++)
+				f2fs_wait_on_block_writeback(sbi, blkaddr);

Do we need this? IIRC, DIO would give create=1.

Yes, we need it. When we are overwriting an existing file, DIO calls f2fs_map_blocks() with create=0. From the DIO code, I see that this happens because blockdev_direct_IO() passes this dio flag DIO_SKIP_HOLES. And then in get_more_blocks(), below code updates create=0, when we are overwriting an existing file.

                create = dio->op == REQ_OP_WRITE;
                if (dio->flags & DIO_SKIP_HOLES) {
                        if (fs_startblk <= ((i_size_read(dio->inode) - 1) >>
                                                        i_blkbits))
                                create = 0;
                }

Quoted 7 lines

Quoted 3 lines

                ret = (*sdio->get_block)(dio->inode, fs_startblk,
                                                map_bh, create);

Got it.
How about this?

Sorry, this is v2.

Quoted 1 line

From b78dd7b2e0317be18716b9496269e9792829f63e Mon Sep 17 00:00:00 2001

From: Sahitya Tummala <[email protected]>
Date: Wed, 10 Oct 2018 10:56:22 +0530
Subject: [PATCH] f2fs: fix data corruption issue with hardware encryption

Direct IO can be used in case of hardware encryption. The following scenario results into data corruption issue in this path -

Thread A - Thread B-
-> write file#1 in direct IO

                                    -> GC gets kicked in
                                    -> GC submitted bio on meta mapping
				       for file#1, but pending completion
-> write file#1 again with new data
   in direct IO
                                    -> GC bio gets completed now
                                    -> GC writes old data to the new
                                       location and thus file#1 is
				       corrupted.

Fix this by submitting and waiting for pending io on meta mapping for direct IO case in f2fs_map_blocks().

Signed-off-by Sahitya Tummala <[email protected]>

Signed-off-by Jaegeuk Kim <[email protected]>