On 2020/6/24 9:21, Jaegeuk Kim wrote: > If two readahead threads having same offset enter in readpages, every read > IOs are split and issued to the disk which giving lower bandwidth. > > This patch tries to avoid redundant readahead calls. > > Signed-off-by: Jaegeuk Kim <jaeg...@kernel.org> > --- > fs/f2fs/data.c | 15 +++++++++++++++ > fs/f2fs/f2fs.h | 1 + > fs/f2fs/super.c | 2 ++ > 3 files changed, 18 insertions(+) > > diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c > index dfd3225153570..1886d83bc5f15 100644 > --- a/fs/f2fs/data.c > +++ b/fs/f2fs/data.c > @@ -2292,6 +2292,7 @@ static int f2fs_mpage_readpages(struct inode *inode, > unsigned nr_pages = rac ? readahead_count(rac) : 1; > unsigned max_nr_pages = nr_pages; > int ret = 0; > + bool drop_ra = false; > > map.m_pblk = 0; > map.m_lblk = 0; > @@ -2302,6 +2303,17 @@ static int f2fs_mpage_readpages(struct inode *inode, > map.m_seg_type = NO_CHECK_TYPE; > map.m_may_create = false; > > + /* > + * Two readahead threads for same address range can cause race condition > + * which fragments sequential read IOs. So let's avoid each other. > + */ > + if (rac && readahead_count(rac)) { > + if (F2FS_I(inode)->ra_offset == readahead_index(rac)) > + drop_ra = true;
I guess you missed to return at somewhere when drop_ra is true? thanks, > + else > + F2FS_I(inode)->ra_offset = readahead_index(rac); > + } > + > for (; nr_pages; nr_pages--) { > if (rac) { > page = readahead_page(rac); > @@ -2368,6 +2380,9 @@ static int f2fs_mpage_readpages(struct inode *inode, > } > if (bio) > __submit_bio(F2FS_I_SB(inode), bio, DATA); > + > + if (rac && readahead_count(rac) && !drop_ra) > + F2FS_I(inode)->ra_offset = -1; > return ret; > } > > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h > index 7fb2a1a334388..753782426feac 100644 > --- a/fs/f2fs/f2fs.h > +++ b/fs/f2fs/f2fs.h > @@ -809,6 +809,7 @@ struct f2fs_inode_info { > struct list_head inmem_pages; /* inmemory pages managed by f2fs */ > struct task_struct *inmem_task; /* store inmemory task */ > struct mutex inmem_lock; /* lock for inmemory pages */ > + pgoff_t ra_offset; /* ongoing readahead offset */ > struct extent_tree *extent_tree; /* cached extent_tree entry */ > > /* avoid racing between foreground op and gc */ > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > index 7326522057378..80cb7cd358f84 100644 > --- a/fs/f2fs/super.c > +++ b/fs/f2fs/super.c > @@ -1015,6 +1015,8 @@ static struct inode *f2fs_alloc_inode(struct > super_block *sb) > /* Will be used by directory only */ > fi->i_dir_level = F2FS_SB(sb)->dir_level; > > + fi->ra_offset = -1; > + > return &fi->vfs_inode; > } > > _______________________________________________ Linux-f2fs-devel mailing list Linux-f2fs-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel