f2fs: allocate blocks for pinned file
Subsystems (2)
Jaegeuk Kim
maintainerPATCH 1/1 f2fs: allocate blocks for pinned file
This patch allows fallocate to allocate physical blocks for pinned file.
--- fs/f2fs/file.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
fs/f2fs/file.c +6 -1
@@ -1528,7 +1528,12 @@ static int expand_inode_data(struct inode *inode, loff_t offset,if (off_end) map.m_len++; - err =f2fs_map_blocks (inode, &map,1 ,F2FS_GET_BLOCK_PRE_AIO ); +if (f2fs_is_pinned_file (inode)) + map.m_seg_type =CURSEG_COLD_DATA ; + + err =f2fs_map_blocks (inode, &map,1 , (f2fs_is_pinned_file (inode) ? +F2FS_GET_BLOCK_PRE_DIO : +F2FS_GET_BLOCK_PRE_AIO ));if (err) {pgoff_t last_off;
Signature · 1 line
2.19.0.605.g01d371f741-goog
Hi Jaegeuk,
Quoted 1 line · Jaegeuk Kim
This patch allows fallocate to allocate physical blocks for pinned file.
Quoted from manual of fallocate(2):
"
Any subregion within the range specified by offset and len that did not contain data before the call will be initialized to zero.
If the FALLOC_FL_KEEP_SIZE flag is specified in mode .... Preallocating zeroed blocks beyond the end of the file in this manner is useful for optimizing append workloads.
"
As quoted description, our change may break the rule of fallocate(, mode = 0), because with after this change, we can't guarantee that preallocated physical block contains zeroed data
Should we introduce an additional ioctl for this case? Or maybe add one more flag in fallocate() for unzeroed block preallocation, not sure.
Thanks,
Quoted 25 lines
Signed-off-by Jaegeuk Kim <[email protected]>--- fs/f2fs/file.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)fs/f2fs/file.c +6
-1@@ -1528,7 +1528,12 @@ static int expand_inode_data(struct inode *inode, loff_t offset,if (off_end) map.m_len++; - err =f2fs_map_blocks (inode, &map,1 ,F2FS_GET_BLOCK_PRE_AIO ); +if (f2fs_is_pinned_file (inode)) + map.m_seg_type =CURSEG_COLD_DATA ; + + err =f2fs_map_blocks (inode, &map,1 , (f2fs_is_pinned_file (inode) ? +F2FS_GET_BLOCK_PRE_DIO : +F2FS_GET_BLOCK_PRE_AIO ));if (err) {pgoff_t last_off;
Jaegeuk Kim
maintainerQuoted 15 lines · Chao Yu
Hi Jaegeuk,
Quoted 1 line · Jaegeuk Kim
This patch allows fallocate to allocate physical blocks for pinned file.
Quoted from manual of fallocate(2):
"
Any subregion within the range specified by offset and len that did not contain data before the call will be initialized to zero.If the
FALLOC_FL_KEEP_SIZEflag is specified in mode .... Preallocating zeroed blocks beyond the end of the file in this manner is useful for optimizing append workloads.
"
Quoted 6 lines
As quoted description, our change may break the rule of
fallocate(, mode = 0), because with after this change, we can't guarantee that preallocated physical block contains zeroed dataShould we introduce an additional ioctl for this case? Or maybe add one more flag in
fallocate()for unzeroed block preallocation, not sure.
I thought like that, but this is a very corner case for the pinned file only in f2fs. And, the pinned file is also indeed used by this purpose.
Thanks,
Quoted 28 lines
Thanks,
Quoted 25 lines
Signed-off-by Jaegeuk Kim <[email protected]>--- fs/f2fs/file.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)fs/f2fs/file.c +6
-1@@ -1528,7 +1528,12 @@ static int expand_inode_data(struct inode *inode, loff_t offset,if (off_end) map.m_len++; - err =f2fs_map_blocks (inode, &map,1 ,F2FS_GET_BLOCK_PRE_AIO ); +if (f2fs_is_pinned_file (inode)) + map.m_seg_type =CURSEG_COLD_DATA ; + + err =f2fs_map_blocks (inode, &map,1 , (f2fs_is_pinned_file (inode) ? +F2FS_GET_BLOCK_PRE_DIO : +F2FS_GET_BLOCK_PRE_AIO ));if (err) {pgoff_t last_off;
Quoted 20 lines · Jaegeuk Kim
Quoted 19 lines · Chao Yu
Hi Jaegeuk,
Quoted 1 line · Jaegeuk Kim
This patch allows fallocate to allocate physical blocks for pinned file.
Quoted from manual of fallocate(2):
"
Any subregion within the range specified by offset and len that did not contain data before the call will be initialized to zero.If the
FALLOC_FL_KEEP_SIZEflag is specified in mode .... Preallocating zeroed blocks beyond the end of the file in this manner is useful for optimizing append workloads.
"As quoted description, our change may break the rule of
fallocate(, mode = 0), because with after this change, we can't guarantee that preallocated physical block contains zeroed data
Quoted 5 lines
Quoted 2 lines
Should we introduce an additional ioctl for this case? Or maybe add one more flag in
fallocate()for unzeroed block preallocation, not sure.I thought like that, but this is a very corner case for the pinned file only in f2fs. And, the pinned file is also indeed used by this purpose.
Okay, I think we need to find one place to document such behavior that is not matching regular poxis fallocate's sematic, user should be noticed about it.
Thanks,
Quoted 33 lines
Thanks,
Quoted 17 lines
Thanks,
Quoted 14 lines
Signed-off-by Jaegeuk Kim <[email protected]>--- fs/f2fs/file.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)fs/f2fs/file.c +0
-0@@ -1528,7 +1528,12 @@ static int expand_inode_data(struct inode *inode, loff_t offset,if (off_end) map.m_len++;Quoted 11 lines
fs/f2fs/file.c static int expand_inode_data(struct inode *inode, loff_t offset, · line 1539
Quoted 11 lines
fs/f2fs/file.c +6
-1- err =f2fs_map_blocks (inode, &map,1 ,F2FS_GET_BLOCK_PRE_AIO ); +if (f2fs_is_pinned_file (inode)) + map.m_seg_type =CURSEG_COLD_DATA ; + + err =f2fs_map_blocks (inode, &map,1 , (f2fs_is_pinned_file (inode) ? +F2FS_GET_BLOCK_PRE_DIO : +F2FS_GET_BLOCK_PRE_AIO ));if (err) {pgoff_t last_off;.
Jaegeuk Kim
maintainerThis patch allows fallocate to allocate physical blocks for pinned file.
--- Documentation/filesystems/f2fs.txt | 25 +++++++++++++++++++++++++ fs/f2fs/file.c | 7 ++++++- 2 files changed, 31 insertions(+), 1 deletion(-)
Documentation/filesystems/f2fs.txt +25 -0
@@ -804,3 +804,28 @@ WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET WRITE_LIFE_NONE " WRITE_LIFE_NONE WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM WRITE_LIFE_LONG " WRITE_LIFE_LONG + +Fallocate(2) Policy +------------------- + +The default policy follows the below posix rule. + +Allocating disk space + The default operation (i.e., mode is zero) of fallocate() allocates + the disk space within the range specified by offset and len. The + file size (as reported by stat(2)) will be changed if offset+len is + greater than the file size. Any subregion within the range specified + by offset and len that did not contain data before the call will be + initialized to zero. This default behavior closely resembles the + behavior of the posix_fallocate(3) library function, and is intended + as a method of optimally implementing that function. + +However, once F2FS receives ioctl(fd, F2FS_IOC_SET_PIN_FILE) in prior to +fallocate(fd, DEFAULT_MODE), it allocates on-disk blocks addressess having +zero or random data, which is useful to the below scenario where: + 1. create(fd) + 2. ioctl(fd, F2FS_IOC_SET_PIN_FILE) + 3. fallocate(fd, 0, 0, size) + 4. address = fibmap(fd, offset) + 5. open(blkdev) + 6. write(blkdev, address)
fs/f2fs/file.c +6 -1
@@ -1528,7 +1528,12 @@ static int expand_inode_data(struct inode *inode, loff_t offset,if (off_end) map.m_len++; - err =f2fs_map_blocks (inode, &map,1 ,F2FS_GET_BLOCK_PRE_AIO ); +if (f2fs_is_pinned_file (inode)) + map.m_seg_type =CURSEG_COLD_DATA ; + + err =f2fs_map_blocks (inode, &map,1 , (f2fs_is_pinned_file (inode) ? +F2FS_GET_BLOCK_PRE_DIO : +F2FS_GET_BLOCK_PRE_AIO ));if (err) {pgoff_t last_off;
Signature · 1 line
2.19.0.605.g01d371f741-goog
Chao Yu
Reviewed-byQuoted 3 lines · Jaegeuk Kim
This patch allows fallocate to allocate physical blocks for pinned file.
Signed-off-by Jaegeuk Kim <[email protected]>
Looks good to me now.
Thanks,
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel