| Message ID | 20240223205535.307307-2-jaegeuk@kernel.org (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [f2fs-dev,1/5] f2fs: check number of blocks in a current section | expand |
On 2024/2/24 4:55, Jaegeuk Kim wrote: > Even if the roll forward recovery stopped due to any error, we have to fix > the write pointers in order to mount the disk from the previous checkpoint. Jaegeuk, IIUC, we may lost warm node chain once we allocate new section for all logs, should we give some notification in log to indicate such condition that filesystem doesn't process a full recovery flow? Thanks, > > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> > --- > fs/f2fs/recovery.c | 2 +- > fs/f2fs/super.c | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c > index b3baec666afe..8bbecb5f9323 100644 > --- a/fs/f2fs/recovery.c > +++ b/fs/f2fs/recovery.c > @@ -913,7 +913,7 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) > * and the f2fs is not read only, check and fix zoned block devices' > * write pointer consistency. > */ > - if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && > + if (fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && > f2fs_sb_has_blkzoned(sbi)) { > err = f2fs_fix_curseg_write_pointer(sbi); > if (!err) > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > index 2e41142d07c0..4d03ce1109ad 100644 > --- a/fs/f2fs/super.c > +++ b/fs/f2fs/super.c > @@ -4673,7 +4673,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) > * If the f2fs is not readonly and fsync data recovery succeeds, > * check zoned block devices' write pointer consistency. > */ > - if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { > + if (!f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { > err = f2fs_check_write_pointer(sbi); > if (err) > goto free_meta;
On Sun, Feb 25, 2024 at 6:49 PM Chao Yu <chao@kernel.org> wrote: > > On 2024/2/24 4:55, Jaegeuk Kim wrote: > > Even if the roll forward recovery stopped due to any error, we have to fix > > the write pointers in order to mount the disk from the previous checkpoint. > > Jaegeuk, > > IIUC, we may lost warm node chain once we allocate new section for all logs, > should we give some notification in log to indicate such condition that > filesystem doesn't process a full recovery flow? > > Thanks, > > > > > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> > > --- > > fs/f2fs/recovery.c | 2 +- > > fs/f2fs/super.c | 2 +- > > 2 files changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c > > index b3baec666afe..8bbecb5f9323 100644 > > --- a/fs/f2fs/recovery.c > > +++ b/fs/f2fs/recovery.c > > @@ -913,7 +913,7 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) > > * and the f2fs is not read only, check and fix zoned block devices' > > * write pointer consistency. > > */ > > - if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && > > + if (fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && > > f2fs_sb_has_blkzoned(sbi)) { > > err = f2fs_fix_curseg_write_pointer(sbi); We might overwrite the previous err value here with a new err from f2fs_fix_curseg_write_pointer(). > > if (!err) > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > > index 2e41142d07c0..4d03ce1109ad 100644 > > --- a/fs/f2fs/super.c > > +++ b/fs/f2fs/super.c > > @@ -4673,7 +4673,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) > > * If the f2fs is not readonly and fsync data recovery succeeds, > > * check zoned block devices' write pointer consistency. > > */ > > - if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { > > + if (!f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { > > err = f2fs_check_write_pointer(sbi); ditto > > if (err) > > goto free_meta; > > > _______________________________________________ > Linux-f2fs-devel mailing list > Linux-f2fs-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Even if the roll forward recovery stopped due to any error, we have to fix
the write pointers in order to mount the disk from the previous checkpoint.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
from v1:
- preserve error
fs/f2fs/recovery.c | 15 +++++++--------
fs/f2fs/super.c | 11 +++++++----
2 files changed, 14 insertions(+), 12 deletions(-)
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
index b3baec666afe..3078d5613748 100644
--- a/fs/f2fs/recovery.c
+++ b/fs/f2fs/recovery.c
@@ -863,7 +863,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
int ret = 0;
unsigned long s_flags = sbi->sb->s_flags;
bool need_writecp = false;
- bool fix_curseg_write_pointer = false;
if (is_sbi_flag_set(sbi, SBI_IS_WRITABLE))
f2fs_info(sbi, "recover fsync data on readonly fs");
@@ -894,8 +893,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
else
f2fs_bug_on(sbi, sbi->sb->s_flags & SB_ACTIVE);
skip:
- fix_curseg_write_pointer = !check_only || list_empty(&inode_list);
-
destroy_fsync_dnodes(&inode_list, err);
destroy_fsync_dnodes(&tmp_inode_list, err);
@@ -913,11 +910,13 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only)
* and the f2fs is not read only, check and fix zoned block devices'
* write pointer consistency.
*/
- if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) &&
- f2fs_sb_has_blkzoned(sbi)) {
- err = f2fs_fix_curseg_write_pointer(sbi);
- if (!err)
- err = f2fs_check_write_pointer(sbi);
+ if (f2fs_sb_has_blkzoned(sbi) && !f2fs_readonly(sbi->sb)) {
+ int err2 = f2fs_fix_curseg_write_pointer(sbi);
+
+ if (!err2)
+ err2 = f2fs_check_write_pointer(sbi);
+ if (err2)
+ err = err2;
ret = err;
}
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index d91e57ca6110..77348fd0a42b 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -4674,11 +4674,14 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
* If the f2fs is not readonly and fsync data recovery succeeds,
* check zoned block devices' write pointer consistency.
*/
- if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) {
- err = f2fs_check_write_pointer(sbi);
- if (err)
- goto free_meta;
+ if (f2fs_sb_has_blkzoned(sbi) && !f2fs_readonly(sb)) {
+ int err2 = f2fs_check_write_pointer(sbi);
+
+ if (err2)
+ err = err2;
}
+ if (err)
+ goto free_meta;
f2fs_init_inmem_curseg(sbi);
On 02/26, Chao Yu wrote: > On 2024/2/24 4:55, Jaegeuk Kim wrote: > > Even if the roll forward recovery stopped due to any error, we have to fix > > the write pointers in order to mount the disk from the previous checkpoint. > > Jaegeuk, > > IIUC, we may lost warm node chain once we allocate new section for all logs, > should we give some notification in log to indicate such condition that > filesystem doesn't process a full recovery flow? How about v2 to preserve the error which gives a warnings on roll-forward recovery? > > Thanks, > > > > > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> > > --- > > fs/f2fs/recovery.c | 2 +- > > fs/f2fs/super.c | 2 +- > > 2 files changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c > > index b3baec666afe..8bbecb5f9323 100644 > > --- a/fs/f2fs/recovery.c > > +++ b/fs/f2fs/recovery.c > > @@ -913,7 +913,7 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) > > * and the f2fs is not read only, check and fix zoned block devices' > > * write pointer consistency. > > */ > > - if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && > > + if (fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && > > f2fs_sb_has_blkzoned(sbi)) { > > err = f2fs_fix_curseg_write_pointer(sbi); > > if (!err) > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > > index 2e41142d07c0..4d03ce1109ad 100644 > > --- a/fs/f2fs/super.c > > +++ b/fs/f2fs/super.c > > @@ -4673,7 +4673,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) > > * If the f2fs is not readonly and fsync data recovery succeeds, > > * check zoned block devices' write pointer consistency. > > */ > > - if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { > > + if (!f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { > > err = f2fs_check_write_pointer(sbi); > > if (err) > > goto free_meta;
On 2024/2/27 8:59, Jaegeuk Kim wrote: > Even if the roll forward recovery stopped due to any error, we have to fix > the write pointers in order to mount the disk from the previous checkpoint. > > Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> Reviewed-by: Chao Yu <chao@kernel.org> Thanks,
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index b3baec666afe..8bbecb5f9323 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -913,7 +913,7 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) * and the f2fs is not read only, check and fix zoned block devices' * write pointer consistency. */ - if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && + if (fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && f2fs_sb_has_blkzoned(sbi)) { err = f2fs_fix_curseg_write_pointer(sbi); if (!err) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 2e41142d07c0..4d03ce1109ad 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -4673,7 +4673,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) * If the f2fs is not readonly and fsync data recovery succeeds, * check zoned block devices' write pointer consistency. */ - if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { + if (!f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { err = f2fs_check_write_pointer(sbi); if (err) goto free_meta;
Even if the roll forward recovery stopped due to any error, we have to fix the write pointers in order to mount the disk from the previous checkpoint. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> --- fs/f2fs/recovery.c | 2 +- fs/f2fs/super.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)