f2fs: support F2FS_IOC_PRECACHE_EXTENTS

This patch introduces a new ioctl F2FS_IOC_PRECACHE_EXTENTS to precache
extent info like ext4, in order to gain better performance during
triggering AIO by eliminating synchronous waiting of mapping info.

Referred commit: 7869a4a6c5ca ("ext4: add support for extent pre-caching")

In addition, with newly added extent precache abilitiy, this patch add
to support FIEMAP_FLAG_CACHE in ->fiemap.

Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 945f15607484ca68baf3c0da93fdd8e7f5c1d895..28a3328f161700831680addc267a4ad18dedf6d8 100644 (file)
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -865,6 +865,7 @@ int f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *from)
                map.m_len = 0;
 
        map.m_next_pgofs = NULL;
+       map.m_next_extent = NULL;
        map.m_seg_type = NO_CHECK_TYPE;
 
        if (direct_io) {
@@ -931,6 +932,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
        blkcnt_t prealloc;
        struct extent_info ei = {0,0,0};
        block_t blkaddr;
+       unsigned int start_pgofs;
 
        if (!maxblocks)
                return 0;
@@ -946,6 +948,8 @@ 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;
+               if (map->m_next_extent)
+                       *map->m_next_extent = pgofs + map->m_len;
                goto out;
        }
 
@@ -964,10 +968,14 @@ next_dnode:
                        if (map->m_next_pgofs)
                                *map->m_next_pgofs =
                                        get_next_page_offset(&dn, pgofs);
+                       if (map->m_next_extent)
+                               *map->m_next_extent =
+                                       get_next_page_offset(&dn, pgofs);
                }
                goto unlock_out;
        }
 
+       start_pgofs = pgofs;
        prealloc = 0;
        last_ofs_in_node = ofs_in_node = dn.ofs_in_node;
        end_offset = ADDRS_PER_PAGE(dn.node_page, inode);
@@ -1001,6 +1009,8 @@ next_block:
                                map->m_pblk = 0;
                                goto sync_out;
                        }
+                       if (flag == F2FS_GET_BLOCK_PRECACHE)
+                               goto sync_out;
                        if (flag == F2FS_GET_BLOCK_FIEMAP &&
                                                blkaddr == NULL_ADDR) {
                                if (map->m_next_pgofs)
@@ -1059,6 +1069,16 @@ skip:
        else if (dn.ofs_in_node < end_offset)
                goto next_block;
 
+       if (flag == F2FS_GET_BLOCK_PRECACHE) {
+               if (map->m_flags & F2FS_MAP_MAPPED) {
+                       unsigned int ofs = start_pgofs - map->m_lblk;
+
+                       f2fs_update_extent_cache_range(&dn,
+                               start_pgofs, map->m_pblk + ofs,
+                               map->m_len - ofs);
+               }
+       }
+
        f2fs_put_dnode(&dn);
 
        if (create) {
@@ -1068,6 +1088,17 @@ skip:
        goto next_dnode;
 
 sync_out:
+       if (flag == F2FS_GET_BLOCK_PRECACHE) {
+               if (map->m_flags & F2FS_MAP_MAPPED) {
+                       unsigned int ofs = start_pgofs - map->m_lblk;
+
+                       f2fs_update_extent_cache_range(&dn,
+                               start_pgofs, map->m_pblk + ofs,
+                               map->m_len - ofs);
+               }
+               if (map->m_next_extent)
+                       *map->m_next_extent = pgofs + 1;
+       }
        f2fs_put_dnode(&dn);
 unlock_out:
        if (create) {
@@ -1089,6 +1120,7 @@ static int __get_data_block(struct inode *inode, sector_t iblock,
        map.m_lblk = iblock;
        map.m_len = bh->b_size >> inode->i_blkbits;
        map.m_next_pgofs = next_pgofs;
+       map.m_next_extent = NULL;
        map.m_seg_type = seg_type;
 
        err = f2fs_map_blocks(inode, &map, create, flag);
@@ -1212,6 +1244,12 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
        u32 flags = 0;
        int ret = 0;
 
+       if (fieinfo->fi_flags & FIEMAP_FLAG_CACHE) {
+               ret = f2fs_precache_extents(inode);
+               if (ret)
+                       return ret;
+       }
+
        ret = fiemap_check_flags(fieinfo, FIEMAP_FLAG_SYNC | FIEMAP_FLAG_XATTR);
        if (ret)
                return ret;
@@ -1313,6 +1351,7 @@ static int f2fs_mpage_readpages(struct address_space *mapping,
        map.m_len = 0;
        map.m_flags = 0;
        map.m_next_pgofs = NULL;
+       map.m_next_extent = NULL;
        map.m_seg_type = NO_CHECK_TYPE;
 
        for (; nr_pages; nr_pages--) {

diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index d5090b4d2633730f66041b3b63178e852972af40..7a85298a74eb561546d1ab4729922ffb3d293448 100644 (file)
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -358,6 +358,7 @@ static inline bool __has_cursum_space(struct f2fs_journal *journal,
 #define F2FS_IOC_GET_FEATURES          _IOR(F2FS_IOCTL_MAGIC, 12, __u32)
 #define F2FS_IOC_SET_PIN_FILE          _IOW(F2FS_IOCTL_MAGIC, 13, __u32)
 #define F2FS_IOC_GET_PIN_FILE          _IOR(F2FS_IOCTL_MAGIC, 14, __u32)
+#define F2FS_IOC_PRECACHE_EXTENTS      _IO(F2FS_IOCTL_MAGIC, 15)
 
 #define F2FS_IOC_SET_ENCRYPTION_POLICY FS_IOC_SET_ENCRYPTION_POLICY
 #define F2FS_IOC_GET_ENCRYPTION_POLICY FS_IOC_GET_ENCRYPTION_POLICY
@@ -553,6 +554,7 @@ struct f2fs_map_blocks {
        unsigned int m_len;
        unsigned int m_flags;
        pgoff_t *m_next_pgofs;          /* point next possible non-hole pgofs */
+       pgoff_t *m_next_extent;         /* point to next possible extent */
        int m_seg_type;
 };
 
@@ -563,6 +565,7 @@ enum {
        F2FS_GET_BLOCK_BMAP,
        F2FS_GET_BLOCK_PRE_DIO,
        F2FS_GET_BLOCK_PRE_AIO,
+       F2FS_GET_BLOCK_PRECACHE,
 };
 
 /*
@@ -2564,6 +2567,7 @@ int f2fs_getattr(const struct path *path, struct kstat *stat,
 int f2fs_setattr(struct dentry *dentry, struct iattr *attr);
 int truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end);
 void truncate_data_blocks_range(struct dnode_of_data *dn, int count);
+int f2fs_precache_extents(struct inode *inode);
 long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
 long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 int f2fs_pin_file_control(struct inode *inode, bool inc);

diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 493457a200b135f36ae51804672c732056b2db52..58c9e3d8b8ec81a35e0fcad8eea11a6169e06436 100644 (file)
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -1423,7 +1423,7 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
 {
        struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
        struct f2fs_map_blocks map = { .m_next_pgofs = NULL,
-                                       .m_seg_type = NO_CHECK_TYPE };
+                       .m_next_extent = NULL, .m_seg_type = NO_CHECK_TYPE };
        pgoff_t pg_end;
        loff_t new_size = i_size_read(inode);
        loff_t off_end;
@@ -2072,7 +2072,7 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi,
 {
        struct inode *inode = file_inode(filp);
        struct f2fs_map_blocks map = { .m_next_pgofs = NULL,
-                                       .m_seg_type = NO_CHECK_TYPE };
+                       .m_next_extent = NULL, .m_seg_type = NO_CHECK_TYPE };
        struct extent_info ei = {0,0,0};
        pgoff_t pg_start, pg_end;
        unsigned int blk_per_seg = sbi->blocks_per_seg;
@@ -2749,6 +2749,43 @@ static int f2fs_ioc_get_pin_file(struct file *filp, unsigned long arg)
        return put_user(pin, (u32 __user *)arg);
 }
 
+int f2fs_precache_extents(struct inode *inode)
+{
+       struct f2fs_inode_info *fi = F2FS_I(inode);
+       struct f2fs_map_blocks map;
+       pgoff_t m_next_extent;
+       loff_t end;
+       int err;
+
+       if (is_inode_flag_set(inode, FI_NO_EXTENT))
+               return -EOPNOTSUPP;
+
+       map.m_lblk = 0;
+       map.m_next_pgofs = NULL;
+       map.m_next_extent = &m_next_extent;
+       map.m_seg_type = NO_CHECK_TYPE;
+       end = F2FS_I_SB(inode)->max_file_blocks;
+
+       while (map.m_lblk < end) {
+               map.m_len = end - map.m_lblk;
+
+               down_write(&fi->dio_rwsem[WRITE]);
+               err = f2fs_map_blocks(inode, &map, 0, F2FS_GET_BLOCK_PRECACHE);
+               up_write(&fi->dio_rwsem[WRITE]);
+               if (err)
+                       return err;
+
+               map.m_lblk = m_next_extent;
+       }
+
+       return err;
+}
+
+static int f2fs_ioc_precache_extents(struct file *filp, unsigned long arg)
+{
+       return f2fs_precache_extents(file_inode(filp));
+}
+
 long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
        if (unlikely(f2fs_cp_error(F2FS_I_SB(file_inode(filp)))))
@@ -2803,6 +2840,8 @@ long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
                return f2fs_ioc_get_pin_file(filp, arg);
        case F2FS_IOC_SET_PIN_FILE:
                return f2fs_ioc_set_pin_file(filp, arg);
+       case F2FS_IOC_PRECACHE_EXTENTS:
+               return f2fs_ioc_precache_extents(filp, arg);
        default:
                return -ENOTTY;
        }
@@ -2880,6 +2919,7 @@ long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
        case F2FS_IOC_FSSETXATTR:
        case F2FS_IOC_GET_PIN_FILE:
        case F2FS_IOC_SET_PIN_FILE:
+       case F2FS_IOC_PRECACHE_EXTENTS:
                break;
        default:
                return -ENOIOCTLCMD;