On 2018/2/12 11:42, Yunlei He wrote:
> This patch add two bug on check back during flush nat entries,
> nat_tree_lock will prevent race from node block allocation
> in write_begin.
>
> Signed-off-by: Yunlei He <heyun...@huawei.com>
> ---
> fs/f2fs/node.c | 9 ++++-----
> 1 file changed, 4 insertions(+), 5 deletions(-)
>
> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
> index 833b46b..ced986d 100644
> --- a/fs/f2fs/node.c
> +++ b/fs/f2fs/node.c
> @@ -2519,10 +2519,9 @@ static void __flush_nat_entry_set(struct f2fs_sb_info
> *sbi,
> }
>
> /* Allow dirty nats by node block allocation in write_begin */
> - if (!set->entry_cnt) {
> - radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
> - kmem_cache_free(nat_entry_set_slab, set);
During checkpoint, We allow buffered write which needs nid allocation, if we
rollback the codes here, we may encounter the BUG_ON?
Thanks,
> - }
> + f2fs_bug_on(sbi, set->entry_cnt);
> + radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
> + kmem_cache_free(nat_entry_set_slab, set);
> }
>
> /*
> @@ -2566,8 +2565,8 @@ void flush_nat_entries(struct f2fs_sb_info *sbi, struct
> cp_control *cpc)
> list_for_each_entry_safe(set, tmp, &sets, set_list)
> __flush_nat_entry_set(sbi, set, cpc);
>
> + f2fs_bug_on(sbi, nm_i->dirty_nat_cnt);
> up_write(&nm_i->nat_tree_lock);
> - /* Allow dirty nats by node block allocation in write_begin */
> }
>
> static int __get_nat_bitmaps(struct f2fs_sb_info *sbi)
>
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
