Re: [f2fs-dev] [PATCH 2/2 v2] f2fs: avoid f2fs_bug_on if f2fs_get_meta_page_nofail got EIO

From: Chao Yu
Date: Tue Sep 25 2018 - 21:10:14 EST


On 2018/9/26 8:18, Jaegeuk Kim wrote:
> On 09/21, Chao Yu wrote:
>> On 2018/9/21 5:46, Jaegeuk Kim wrote:
>>> On 09/20, Chao Yu wrote:
>>>> On 2018/9/19 1:56, Jaegeuk Kim wrote:
>>>>> This patch avoids BUG_ON when f2fs_get_meta_page_nofail got EIO during
>>>>> xfstests/generic/475.
>>>>>
>>>>> Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
>>>>> ---
>>>>> Change log from v1:
>>>>> - propagate errno
>>>>> - drop sum_pages
>>>>>
>>>>> fs/f2fs/checkpoint.c | 10 +++++-----
>>>>> fs/f2fs/f2fs.h | 2 +-
>>>>> fs/f2fs/gc.c | 9 +++++++++
>>>>> fs/f2fs/node.c | 32 ++++++++++++++++++++++++--------
>>>>> fs/f2fs/recovery.c | 2 ++
>>>>> fs/f2fs/segment.c | 3 +++
>>>>> 6 files changed, 44 insertions(+), 14 deletions(-)
>>>>>
>>>>> diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
>>>>> index 01e0d8f5bbbe..1ddf332ce4b2 100644
>>>>> --- a/fs/f2fs/checkpoint.c
>>>>> +++ b/fs/f2fs/checkpoint.c
>>>>> @@ -119,11 +119,8 @@ struct page *f2fs_get_meta_page_nofail(struct f2fs_sb_info *sbi, pgoff_t index)
>>>>> if (PTR_ERR(page) == -EIO &&
>>>>> ++count <= DEFAULT_RETRY_IO_COUNT)
>>>>> goto retry;
>>>>> -
>>>>> f2fs_stop_checkpoint(sbi, false);
>>>>> - f2fs_bug_on(sbi, 1);
>>>>> }
>>>>> -
>>>>> return page;
>>>>> }
>>>>>
>>>>> @@ -1512,7 +1509,10 @@ int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> ckpt->checkpoint_ver = cpu_to_le64(++ckpt_ver);
>>>>>
>>>>> /* write cached NAT/SIT entries to NAT/SIT area */
>>>>> - f2fs_flush_nat_entries(sbi, cpc);
>>>>> + err = f2fs_flush_nat_entries(sbi, cpc);
>>>>> + if (err)
>>>>> + goto stop;
>>>>
>>>> To make sure, if partial NAT pages become dirty, flush them back to device
>>>> outside checkpoint() is not harmful, right?
>>>
>>> I think so.
>>
>> Oh, one more case, now we use NAT #0, if partial NAT pages were
>> writebacked, then we start to use NAT #1, later, in another checkpoint(),
>> we update those NAT pages again, we will write them to NAT #0, which is
>> belong to last valid checkpoint(), it's may cause corrupted checkpoint in
>> scenario of SPO. So it's harmfull, right?
>
> We already stopped checkpoint anymore, so there'd be no way to proceed another
> checkpoint. Am I missing something?

You're right, we have called f2fs_stop_checkpoint(), so we are safe now. :)

Thanks,

>
>>
>> Thanks,
>>
>>>
>>>>
>>>>> +
>>>>> f2fs_flush_sit_entries(sbi, cpc);
>>>>>
>>>>> /* unlock all the fs_lock[] in do_checkpoint() */
>>>>> @@ -1521,7 +1521,7 @@ int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> f2fs_release_discard_addrs(sbi);
>>>>> else
>>>>> f2fs_clear_prefree_segments(sbi, cpc);
>>>>> -
>>>>> +stop:
>>>>> unblock_operations(sbi);
>>>>> stat_inc_cp_count(sbi->stat_info);
>>>>>
>>>>> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
>>>>> index a0c868127a7c..29021dbc8f2a 100644
>>>>> --- a/fs/f2fs/f2fs.h
>>>>> +++ b/fs/f2fs/f2fs.h
>>>>> @@ -2895,7 +2895,7 @@ int f2fs_recover_xattr_data(struct inode *inode, struct page *page);
>>>>> int f2fs_recover_inode_page(struct f2fs_sb_info *sbi, struct page *page);
>>>>> int f2fs_restore_node_summary(struct f2fs_sb_info *sbi,
>>>>> unsigned int segno, struct f2fs_summary_block *sum);
>>>>> -void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
>>>>> +int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
>>>>> int f2fs_build_node_manager(struct f2fs_sb_info *sbi);
>>>>> void f2fs_destroy_node_manager(struct f2fs_sb_info *sbi);
>>>>> int __init f2fs_create_node_manager_caches(void);
>>>>> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
>>>>> index 4bcc8a59fdef..c7d14282dbf4 100644
>>>>> --- a/fs/f2fs/gc.c
>>>>> +++ b/fs/f2fs/gc.c
>>>>> @@ -1070,6 +1070,15 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi,
>>>>> /* reference all summary page */
>>>>> while (segno < end_segno) {
>>>>> sum_page = f2fs_get_sum_page(sbi, segno++);
>>>>> + if (IS_ERR(sum_page)) {
>>>>> + end_segno = segno - 1;
>>>>> + for (segno = start_segno; segno < end_segno; segno++) {
>>>>> + sum_page = find_get_page(META_MAPPING(sbi),
>>>>> + GET_SUM_BLOCK(sbi, segno));
>>>>> + f2fs_put_page(sum_page, 0);
>>>>
>>>> find_get_page() will add one more reference on page, so we need to call
>>>> f2fs_put_page(sum_page, 0) twice.
>>>
>>> Done.
>>>
>>>>
>>>> Thanks,
>>>>
>>>>> + }
>>>>> + return PTR_ERR(sum_page);
>>>>> + }
>>>>> unlock_page(sum_page);
>>>>> }
>>>>>
>>>>> diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
>>>>> index fa2381c0bc47..79b6fee354f7 100644
>>>>> --- a/fs/f2fs/node.c
>>>>> +++ b/fs/f2fs/node.c
>>>>> @@ -126,6 +126,8 @@ static struct page *get_next_nat_page(struct f2fs_sb_info *sbi, nid_t nid)
>>>>>
>>>>> /* get current nat block page with lock */
>>>>> src_page = get_current_nat_page(sbi, nid);
>>>>> + if (IS_ERR(src_page))
>>>>> + return src_page;
>>>>> dst_page = f2fs_grab_meta_page(sbi, dst_off);
>>>>> f2fs_bug_on(sbi, PageDirty(src_page));
>>>>>
>>>>> @@ -2265,15 +2267,19 @@ static int __f2fs_build_free_nids(struct f2fs_sb_info *sbi,
>>>>> nm_i->nat_block_bitmap)) {
>>>>> struct page *page = get_current_nat_page(sbi, nid);
>>>>>
>>>>> - ret = scan_nat_page(sbi, page, nid);
>>>>> - f2fs_put_page(page, 1);
>>>>> + if (IS_ERR(page)) {
>>>>> + ret = PTR_ERR(page);
>>>>> + } else {
>>>>> + ret = scan_nat_page(sbi, page, nid);
>>>>> + f2fs_put_page(page, 1);
>>>>> + }
>>>>>
>>>>> if (ret) {
>>>>> up_read(&nm_i->nat_tree_lock);
>>>>> f2fs_bug_on(sbi, !mount);
>>>>> f2fs_msg(sbi->sb, KERN_ERR,
>>>>> "NAT is corrupt, run fsck to fix it");
>>>>> - return -EINVAL;
>>>>> + return ret;
>>>>> }
>>>>> }
>>>>>
>>>>> @@ -2700,7 +2706,7 @@ static void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid,
>>>>> __clear_bit_le(nat_index, nm_i->full_nat_bits);
>>>>> }
>>>>>
>>>>> -static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
>>>>> +static int __flush_nat_entry_set(struct f2fs_sb_info *sbi,
>>>>> struct nat_entry_set *set, struct cp_control *cpc)
>>>>> {
>>>>> struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
>>>>> @@ -2724,6 +2730,9 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
>>>>> down_write(&curseg->journal_rwsem);
>>>>> } else {
>>>>> page = get_next_nat_page(sbi, start_nid);
>>>>> + if (IS_ERR(page))
>>>>> + return PTR_ERR(page);
>>>>> +
>>>>> nat_blk = page_address(page);
>>>>> f2fs_bug_on(sbi, !nat_blk);
>>>>> }
>>>>> @@ -2769,12 +2778,13 @@ static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
>>>>> radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
>>>>> kmem_cache_free(nat_entry_set_slab, set);
>>>>> }
>>>>> + return 0;
>>>>> }
>>>>>
>>>>> /*
>>>>> * This function is called during the checkpointing process.
>>>>> */
>>>>> -void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> +int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> {
>>>>> struct f2fs_nm_info *nm_i = NM_I(sbi);
>>>>> struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
>>>>> @@ -2784,6 +2794,7 @@ void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> unsigned int found;
>>>>> nid_t set_idx = 0;
>>>>> LIST_HEAD(sets);
>>>>> + int err = 0;
>>>>>
>>>>> /* during unmount, let's flush nat_bits before checking dirty_nat_cnt */
>>>>> if (enabled_nat_bits(sbi, cpc)) {
>>>>> @@ -2793,7 +2804,7 @@ void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> }
>>>>>
>>>>> if (!nm_i->dirty_nat_cnt)
>>>>> - return;
>>>>> + return 0;
>>>>>
>>>>> down_write(&nm_i->nat_tree_lock);
>>>>>
>>>>> @@ -2816,11 +2827,16 @@ void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
>>>>> }
>>>>>
>>>>> /* flush dirty nats in nat entry set */
>>>>> - list_for_each_entry_safe(set, tmp, &sets, set_list)
>>>>> - __flush_nat_entry_set(sbi, set, cpc);
>>>>> + list_for_each_entry_safe(set, tmp, &sets, set_list) {
>>>>> + err = __flush_nat_entry_set(sbi, set, cpc);
>>>>> + if (err)
>>>>> + break;
>>>>> + }
>>>>>
>>>>> up_write(&nm_i->nat_tree_lock);
>>>>> /* Allow dirty nats by node block allocation in write_begin */
>>>>> +
>>>>> + return err;
>>>>> }
>>>>>
>>>>> static int __get_nat_bitmaps(struct f2fs_sb_info *sbi)
>>>>> diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
>>>>> index 56d34193a74b..ba678efe7cad 100644
>>>>> --- a/fs/f2fs/recovery.c
>>>>> +++ b/fs/f2fs/recovery.c
>>>>> @@ -355,6 +355,8 @@ static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi,
>>>>> }
>>>>>
>>>>> sum_page = f2fs_get_sum_page(sbi, segno);
>>>>> + if (IS_ERR(sum_page))
>>>>> + return PTR_ERR(sum_page);
>>>>> sum_node = (struct f2fs_summary_block *)page_address(sum_page);
>>>>> sum = sum_node->entries[blkoff];
>>>>> f2fs_put_page(sum_page, 1);
>>>>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>>>>> index aa96a371aaf8..ab7386a7e9d3 100644
>>>>> --- a/fs/f2fs/segment.c
>>>>> +++ b/fs/f2fs/segment.c
>>>>> @@ -2487,6 +2487,7 @@ static void change_curseg(struct f2fs_sb_info *sbi, int type)
>>>>> __next_free_blkoff(sbi, curseg, 0);
>>>>>
>>>>> sum_page = f2fs_get_sum_page(sbi, new_segno);
>>>>> + f2fs_bug_on(sbi, IS_ERR(sum_page));
>>>>> sum_node = (struct f2fs_summary_block *)page_address(sum_page);
>>>>> memcpy(curseg->sum_blk, sum_node, SUM_ENTRY_SIZE);
>>>>> f2fs_put_page(sum_page, 1);
>>>>> @@ -3971,6 +3972,8 @@ static int build_sit_entries(struct f2fs_sb_info *sbi)
>>>>>
>>>>> se = &sit_i->sentries[start];
>>>>> page = get_current_sit_page(sbi, start);
>>>>> + if (IS_ERR(page))
>>>>> + return PTR_ERR(page);
>>>>> sit_blk = (struct f2fs_sit_block *)page_address(page);
>>>>> sit = sit_blk->entries[SIT_ENTRY_OFFSET(sit_i, start)];
>>>>> f2fs_put_page(page, 1);
>>>>>
>>>
>>> .
>>>
>
> .
>