Skip to contentSkip to conversation

btrfs: fix iomap_begin length for nocow writes

PATCH In review Christoph Hellwig 3 participants 3 messages Jun 8, 2023 – Jun 9, 2023 Btrfs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 btrfs: fix iomap_begin length for nocow writes

can_nocow_extent can reduce the len passed in, which needs to be propagated to btrfs_dio_iomap_begin so that iomap does not submit more data then is mapped.

This problems exists since the btrfs_get_blocks_direct helper was added in commit c5794e51784a ("btrfs: Factor out write portion of btrfs_get_blocks_direct"), but the ordered_extent splitting added in commit b73a6fd1b1ef ("btrfs: split partial dio bios before submit") added a WARN_ON that made a syzcaller test fail.

Fixes c5794e51784a ("btrfs: Factor out write portion of btrfs_get_blocks_direct")
Reported-by [email protected]
---
 fs/btrfs/inode.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)
fs/btrfs/inode.c +12 -6
index 19c707bc8801a9..3f99f02dc1fe20 100644
@@ -7264,7 +7264,7 @@ static struct extent_map *create_io_em(struct btrfs_inode *inode, u64 start,
 static int btrfs_get_blocks_direct_write(struct extent_map **map,
 					 struct inode *inode,
 					 struct btrfs_dio_data *dio_data,
-					 u64 start, u64 len,
+					 u64 start, u64 *lenp,
 					 unsigned int iomap_flags)
 {
 	const bool nowait = (iomap_flags & IOMAP_NOWAIT);
@@ -7275,6 +7275,7 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
 	struct btrfs_block_group *bg;
 	bool can_nocow = false;
 	bool space_reserved = false;
+	u64 len = *lenp;
 	u64 prev_len;
 	int ret = 0;
 
@@ -7345,15 +7346,19 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
 		free_extent_map(em);
 		*map = NULL;
 
-		if (nowait)
-			return -EAGAIN;
+		if (nowait) {
+			ret = -EAGAIN;
+			goto out;
+		}
 
 		/*
 		 * If we could not allocate data space before locking the file
 		 * range and we can't do a NOCOW write, then we have to fail.
 		 */
-		if (!dio_data->data_space_reserved)
-			return -ENOSPC;
+		if (!dio_data->data_space_reserved) {
+			ret = -ENOSPC;
+			goto out;
+		}
 
 		/*
 		 * We have to COW and we have already reserved data space before,
@@ -7394,6 +7399,7 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
 		btrfs_delalloc_release_extents(BTRFS_I(inode), len);
 		btrfs_delalloc_release_metadata(BTRFS_I(inode), len, true);
 	}
+	*lenp = len;
 	return ret;
 }
 
@@ -7570,7 +7576,7 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start,
 
 	if (write) {
 		ret = btrfs_get_blocks_direct_write(&em, inode, dio_data,
-						    start, len, flags);
+						    start, &len, flags);
 		if (ret < 0)
 			goto unlock_err;
 		unlock_extents = true;
Signature · 1 line
2.39.2
Reviewed-by on v1 1/1
#1 lore raw
Quoted 11 lines · Christoph Hellwig

can_nocow_extent can reduce the len passed in, which needs to be propagated to btrfs_dio_iomap_begin so that iomap does not submit more data then is mapped.

This problems exists since the btrfs_get_blocks_direct helper was added in commit c5794e51784a ("btrfs: Factor out write portion of btrfs_get_blocks_direct"), but the ordered_extent splitting added in commit b73a6fd1b1ef ("btrfs: split partial dio bios before submit") added a WARN_ON that made a syzcaller test fail.

Quoted 4 lines
Fixes c5794e51784a ("btrfs: Factor out write portion of btrfs_get_blocks_direct")
Reported-by [email protected]

Looks good to me, thanks.

Quoted 69 lines
---
 fs/btrfs/inode.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)
fs/btrfs/inode.c +12 -7
index 19c707bc8801a9..3f99f02dc1fe20 100644
@@ -7264,7 +7264,7 @@ static struct extent_map *create_io_em(struct btrfs_inode *inode, u64 start,
 static int btrfs_get_blocks_direct_write(struct extent_map **map,
                                         struct inode *inode,
                                         struct btrfs_dio_data *dio_data,
-                                        u64 start, u64 len,
+                                        u64 start, u64 *lenp,
                                         unsigned int iomap_flags)
 {
        const bool nowait = (iomap_flags & IOMAP_NOWAIT);
@@ -7275,6 +7275,7 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
        struct btrfs_block_group *bg;
        bool can_nocow = false;
        bool space_reserved = false;
+       u64 len = *lenp;
        u64 prev_len;
        int ret = 0;
 
@@ -7345,15 +7346,19 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
                free_extent_map(em);
                *map = NULL;
 
-               if (nowait)
-                       return -EAGAIN;
+               if (nowait) {
+                       ret = -EAGAIN;
+                       goto out;
+               }
 
                /*
                 * If we could not allocate data space before locking the file
                 * range and we can't do a NOCOW write, then we have to fail.
                 */
-               if (!dio_data->data_space_reserved)
-                       return -ENOSPC;
+               if (!dio_data->data_space_reserved) {
+                       ret = -ENOSPC;
+                       goto out;
+               }
 
                /*
                 * We have to COW and we have already reserved data space before,
@@ -7394,6 +7399,7 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
                btrfs_delalloc_release_extents(BTRFS_I(inode), len);
                btrfs_delalloc_release_metadata(BTRFS_I(inode), len, true);
        }
+       *lenp = len;
        return ret;
 }
 
@@ -7570,7 +7576,7 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start,
 
        if (write) {
                ret = btrfs_get_blocks_direct_write(&em, inode, dio_data,
-                                                   start, len, flags);
+                                                   start, &len, flags);
                if (ret < 0)
                        goto unlock_err;
                unlock_extents = true;
--

2.39.2

#2 lore raw
In reply to Christoph Hellwig
Quoted 10 lines · Christoph Hellwig

can_nocow_extent can reduce the len passed in, which needs to be propagated to btrfs_dio_iomap_begin so that iomap does not submit more data then is mapped.

This problems exists since the btrfs_get_blocks_direct helper was added in commit c5794e51784a ("btrfs: Factor out write portion of btrfs_get_blocks_direct"), but the ordered_extent splitting added in commit b73a6fd1b1ef ("btrfs: split partial dio bios before submit") added a WARN_ON that made a syzcaller test fail.

Quoted 4 lines
Fixes c5794e51784a ("btrfs: Factor out write portion of btrfs_get_blocks_direct")
Reported-by [email protected]

Added to misc-next, thanks.