On Sun, Aug 09, 2009 at 11:23:54PM -0400, Theodore Ts'o wrote:
> The logic around sbi->s_mb_last_group and sbi->s_mb_last_start was all
> screwed up. These fields were getting unconditionally all the time,
> set even when stream allocation had not taken place, and if they were
> being used when the file was smaller than s_mb_stream_request, which
> is when the allocation should _not_ be doing stream allocation.
>
> Fix this by determining whether or not we stream allocation should
> take place once, in ext4_mb_group_or_file(), and setting a flag which
> gets used in ext4_mb_regular_allocator() and ext4_mb_use_best_found().
> This simplifies the code and assures that we are consistently using
> (or not using) the stream allocation logic.
>
> Signed-off-by: "Theodore Ts'o" <tytso@xxxxxxx>
> ---
> fs/ext4/ext4.h | 2 ++
> fs/ext4/mballoc.c | 23 ++++++++++-------------
> 2 files changed, 12 insertions(+), 13 deletions(-)
>
> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> index e267727..70aa951 100644
> --- a/fs/ext4/ext4.h
> +++ b/fs/ext4/ext4.h
> @@ -88,6 +88,8 @@ typedef unsigned int ext4_group_t;
> #define EXT4_MB_HINT_TRY_GOAL 0x0200
> /* blocks already pre-reserved by delayed allocation */
> #define EXT4_MB_DELALLOC_RESERVED 0x0400
> +/* We are doing stream allocation */
> +#define EXT4_MB_STREAM_ALLOC 0x0800
>
>
> struct ext4_allocation_request {
> diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
> index f510a58..a103cb0 100644
> --- a/fs/ext4/mballoc.c
> +++ b/fs/ext4/mballoc.c
> @@ -1361,7 +1361,7 @@ static void ext4_mb_use_best_found(struct ext4_allocation_context *ac,
> ac->alloc_semp = e4b->alloc_semp;
> e4b->alloc_semp = NULL;
> /* store last allocated for subsequent stream allocation */
> - if ((ac->ac_flags & EXT4_MB_HINT_DATA)) {
> + if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) {
> spin_lock(&sbi->s_md_lock);
> sbi->s_mb_last_group = ac->ac_f_ex.fe_group;
> sbi->s_mb_last_start = ac->ac_f_ex.fe_start;
> @@ -1939,7 +1939,6 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac)
> struct ext4_sb_info *sbi;
> struct super_block *sb;
> struct ext4_buddy e4b;
> - loff_t size, isize;
>
> sb = ac->ac_sb;
> sbi = EXT4_SB(sb);
> @@ -1975,20 +1974,16 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac)
> }
>
> bsbits = ac->ac_sb->s_blocksize_bits;
> - /* if stream allocation is enabled, use global goal */
> - size = ac->ac_o_ex.fe_logical + ac->ac_o_ex.fe_len;
> - isize = i_size_read(ac->ac_inode) >> bsbits;
> - if (size < isize)
> - size = isize;
>
> - if (size < sbi->s_mb_stream_request &&
> - (ac->ac_flags & EXT4_MB_HINT_DATA)) {
> + /* if stream allocation is enabled, use global goal */
> + if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) {
> /* TBD: may be hot point */
> spin_lock(&sbi->s_md_lock);
> ac->ac_g_ex.fe_group = sbi->s_mb_last_group;
> ac->ac_g_ex.fe_start = sbi->s_mb_last_start;
> spin_unlock(&sbi->s_md_lock);
> }
> +
> /* Let's just scan groups to find more-less suitable blocks */
> cr = ac->ac_2order ? 0 : 1;
> /*
> @@ -4192,16 +4187,18 @@ static void ext4_mb_group_or_file(struct ext4_allocation_context *ac)
> if (!(ac->ac_flags & EXT4_MB_HINT_DATA))
> return;
>
> + if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
> + return;
> +
> size = ac->ac_o_ex.fe_logical + ac->ac_o_ex.fe_len;
> isize = i_size_read(ac->ac_inode) >> bsbits;
> size = max(size, isize);
>
> /* don't use group allocation for large files */
> - if (size >= sbi->s_mb_stream_request)
> - return;
> -
> - if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
> + if (size >= sbi->s_mb_stream_request) {
> + ac->ac_flags |= EXT4_MB_STREAM_ALLOC;
> return;
> + }
>
> BUG_ON(ac->ac_lg != NULL);
> /*
NAK.
This would give bad allocation pattern for large files. We should be using global goal only for small files
not for large files. Large files should be using neighbour allocated extent block as the goal, so that
we get contiguous blocks.
-aneesh
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
