[PATCH] fsck.f2fs: fix to set CP_LARGE_NAT_BITMAP_FLAG

[PATCH] fsck.f2fs: fix to set CP_LARGE_NAT_BITMAP_FLAG

[Chao Yu]

During fixing checkpoint, if large_nat_bitmap feature is on, we need to
set CP_LARGE_NAT_BITMAP_FLAG flag in checkpoint.

Signed-off-by: Chao Yu <yuchao0@huawei.com>
---
 fsck/fsck.c  | 2 ++
 fsck/mount.c | 2 ++
 2 files changed, 4 insertions(+)

diff --git a/fsck/fsck.c b/fsck/fsck.c
index a17555c..88ed4d7 100644
--- a/fsck/fsck.c
+++ b/fsck/fsck.c
@@ -2038,6 +2038,8 @@ static void fix_checkpoint(struct f2fs_sb_info *sbi)
 		flags |= CP_TRIMMED_FLAG;
 	if (is_set_ckpt_flags(cp, CP_DISABLED_FLAG))
 		flags |= CP_DISABLED_FLAG;
+	if (is_set_ckpt_flags(cp, CP_LARGE_NAT_BITMAP_FLAG))
+		flags |= CP_LARGE_NAT_BITMAP_FLAG;
 
 	if (flags & CP_UMOUNT_FLAG)
 		cp_blocks = 8;
diff --git a/fsck/mount.c b/fsck/mount.c
index aa64e93..20ecf07 100644
--- a/fsck/mount.c
+++ b/fsck/mount.c
@@ -2365,6 +2365,8 @@ void write_checkpoint(struct f2fs_sb_info *sbi)
 		flags |= CP_TRIMMED_FLAG;
 	if (is_set_ckpt_flags(cp, CP_DISABLED_FLAG))
 		flags |= CP_DISABLED_FLAG;
+	if (is_set_ckpt_flags(cp, CP_LARGE_NAT_BITMAP_FLAG))
+		flags |= CP_LARGE_NAT_BITMAP_FLAG;
 
 	set_cp(free_segment_count, get_free_segments(sbi));
 	set_cp(valid_block_count, sbi->total_valid_block_count);
-- 
2.18.0.rc1

Re: [PATCH] fsck.f2fs: fix to set CP_LARGE_NAT_BITMAP_FLAG

[Ju Hyung Park]

Hi Chao,

The output appears to be similar to the one without this patch,
however, it now actually mounts.

[ 2360.127787] F2FS-fs (loop1): Found nat_bits in checkpoint
[ 2360.189038] F2FS-fs (loop1): orphan cleanup on readonly fs
[ 2360.189284] F2FS-fs (loop1): recover fsync data on readonly fs
[ 2360.189297] F2FS-fs (loop1): Mounted with checkpoint version = 761bfa29

Go to here to look at the new fsck output:
http://arter97.com/f2fs/log_2.txt.xz
(previous one is now log_1.txt.xz)

However, something weird goes on.
Without even touching the image after first fsck run, the consecutive
2nd run prints out A LOT of messages:
http://arter97.com/f2fs/log_3.txt.xz

The image seems to mount normally after on both the 1st and the 2nd fsck run.

Thanks.

On Thu, Apr 18, 2019 at 4:49 PM Chao Yu <yuchao0@huawei.com> wrote:
>
> During fixing checkpoint, if large_nat_bitmap feature is on, we need to
> set CP_LARGE_NAT_BITMAP_FLAG flag in checkpoint.
>
> Signed-off-by: Chao Yu <yuchao0@huawei.com>
> ---
>  fsck/fsck.c  | 2 ++
>  fsck/mount.c | 2 ++
>  2 files changed, 4 insertions(+)
>
> diff --git a/fsck/fsck.c b/fsck/fsck.c
> index a17555c..88ed4d7 100644
> --- a/fsck/fsck.c
> +++ b/fsck/fsck.c
> @@ -2038,6 +2038,8 @@ static void fix_checkpoint(struct f2fs_sb_info *sbi)
>                 flags |= CP_TRIMMED_FLAG;
>         if (is_set_ckpt_flags(cp, CP_DISABLED_FLAG))
>                 flags |= CP_DISABLED_FLAG;
> +       if (is_set_ckpt_flags(cp, CP_LARGE_NAT_BITMAP_FLAG))
> +               flags |= CP_LARGE_NAT_BITMAP_FLAG;
>
>         if (flags & CP_UMOUNT_FLAG)
>                 cp_blocks = 8;
> diff --git a/fsck/mount.c b/fsck/mount.c
> index aa64e93..20ecf07 100644
> --- a/fsck/mount.c
> +++ b/fsck/mount.c
> @@ -2365,6 +2365,8 @@ void write_checkpoint(struct f2fs_sb_info *sbi)
>                 flags |= CP_TRIMMED_FLAG;
>         if (is_set_ckpt_flags(cp, CP_DISABLED_FLAG))
>                 flags |= CP_DISABLED_FLAG;
> +       if (is_set_ckpt_flags(cp, CP_LARGE_NAT_BITMAP_FLAG))
> +               flags |= CP_LARGE_NAT_BITMAP_FLAG;
>
>         set_cp(free_segment_count, get_free_segments(sbi));
>         set_cp(valid_block_count, sbi->total_valid_block_count);
> --
> 2.18.0.rc1
>
>
>
> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2fs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Re: [PATCH] fsck.f2fs: fix to set CP_LARGE_NAT_BITMAP_FLAG

[Chao Yu]

Hi Ju Hyung,

On 2019/4/18 21:14, Ju Hyung Park wrote:
> Hi Chao,
> 
> The output appears to be similar to the one without this patch,
> however, it now actually mounts.
> 
> [ 2360.127787] F2FS-fs (loop1): Found nat_bits in checkpoint
> [ 2360.189038] F2FS-fs (loop1): orphan cleanup on readonly fs
> [ 2360.189284] F2FS-fs (loop1): recover fsync data on readonly fs
> [ 2360.189297] F2FS-fs (loop1): Mounted with checkpoint version = 761bfa29
> 
> Go to here to look at the new fsck output:
> http://arter97.com/f2fs/log_2.txt.xz
> (previous one is now log_1.txt.xz)
> 
> However, something weird goes on.
> Without even touching the image after first fsck run, the consecutive
> 2nd run prints out A LOT of messages:
> http://arter97.com/f2fs/log_3.txt.xz
> 
> The image seems to mount normally after on both the 1st and the 2nd fsck run.

Yes, this is because persistence of chksum in checkpoint will damage sit bitmap.

I've sent two patches, could you please try them?

Thanks

> 
> Thanks.
> 
> On Thu, Apr 18, 2019 at 4:49 PM Chao Yu <yuchao0@huawei.com> wrote:
>>
>> During fixing checkpoint, if large_nat_bitmap feature is on, we need to
>> set CP_LARGE_NAT_BITMAP_FLAG flag in checkpoint.
>>
>> Signed-off-by: Chao Yu <yuchao0@huawei.com>
>> ---
>>  fsck/fsck.c  | 2 ++
>>  fsck/mount.c | 2 ++
>>  2 files changed, 4 insertions(+)
>>
>> diff --git a/fsck/fsck.c b/fsck/fsck.c
>> index a17555c..88ed4d7 100644
>> --- a/fsck/fsck.c
>> +++ b/fsck/fsck.c
>> @@ -2038,6 +2038,8 @@ static void fix_checkpoint(struct f2fs_sb_info *sbi)
>>                 flags |= CP_TRIMMED_FLAG;
>>         if (is_set_ckpt_flags(cp, CP_DISABLED_FLAG))
>>                 flags |= CP_DISABLED_FLAG;
>> +       if (is_set_ckpt_flags(cp, CP_LARGE_NAT_BITMAP_FLAG))
>> +               flags |= CP_LARGE_NAT_BITMAP_FLAG;
>>
>>         if (flags & CP_UMOUNT_FLAG)
>>                 cp_blocks = 8;
>> diff --git a/fsck/mount.c b/fsck/mount.c
>> index aa64e93..20ecf07 100644
>> --- a/fsck/mount.c
>> +++ b/fsck/mount.c
>> @@ -2365,6 +2365,8 @@ void write_checkpoint(struct f2fs_sb_info *sbi)
>>                 flags |= CP_TRIMMED_FLAG;
>>         if (is_set_ckpt_flags(cp, CP_DISABLED_FLAG))
>>                 flags |= CP_DISABLED_FLAG;
>> +       if (is_set_ckpt_flags(cp, CP_LARGE_NAT_BITMAP_FLAG))
>> +               flags |= CP_LARGE_NAT_BITMAP_FLAG;
>>
>>         set_cp(free_segment_count, get_free_segments(sbi));
>>         set_cp(valid_block_count, sbi->total_valid_block_count);
>> --
>> 2.18.0.rc1
>>
>>
>>
>> _______________________________________________
>> Linux-f2fs-devel mailing list
>> Linux-f2fs-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
> .
> 

Re: [PATCH] fsck.f2fs: fix to set CP_LARGE_NAT_BITMAP_FLAG

[Ju Hyung Park]

Hi Chao and Jaegeuk.

On Fri, Apr 19, 2019 at 12:46 PM Chao Yu <yuchao0@huawei.com> wrote:
> I've sent two patches, could you please try them?
>
> Thanks

Thanks for the patches.

So it seems like I need to reformat my f2fs partitions, please confirm
if that's the case.
Also, if these patches break backward-compatibility, wouldn't it be
better to rename the feature and mark the old one deprecated?

I would like to see Jaegeuk's review before mkfs'ing again.

Thanks.

Re: [PATCH] fsck.f2fs: fix to set CP_LARGE_NAT_BITMAP_FLAG

[Chao Yu]

On 2019/4/19 12:36, Ju Hyung Park wrote:
> Hi Chao and Jaegeuk.
> 
> On Fri, Apr 19, 2019 at 12:46 PM Chao Yu <yuchao0@huawei.com> wrote:
>> I've sent two patches, could you please try them?
>>
>> Thanks
> 
> Thanks for the patches.
> 
> So it seems like I need to reformat my f2fs partitions, please confirm
> if that's the case.

I think you can try new fsck to check whether it can fix the corrupted image first?

> Also, if these patches break backward-compatibility, wouldn't it be
> better to rename the feature and mark the old one deprecated?

Yeah, that would be one option, I agree and prefer to listen to Jaegeuk's point
of view.

> 
> I would like to see Jaegeuk's review before mkfs'ing again.
> 
> Thanks.
> .
>