On Mon, Aug 12, 2024 at 12:56 PM Jaegeuk Kim <jaeg...@kernel.org> wrote:
>
> On 08/12, Daeho Jeong wrote:
> > On Thu, Aug 8, 2024 at 4:34 PM Jaegeuk Kim <jaeg...@kernel.org> wrote:
> > >
> > > It'll crach fsck.f2fs. Even, let's keep the write pointers and just 
> > > finish it.
> > >
> > > Signed-off-by: Jaegeuk Kim <jaeg...@kernel.org>
> > > ---
> > >  fsck/fsck.c | 20 ++++++++------------
> > >  1 file changed, 8 insertions(+), 12 deletions(-)
> > >
> > > diff --git a/fsck/fsck.c b/fsck/fsck.c
> > > index a18bee9d0c3b..5fe8ad4ef327 100644
> > > --- a/fsck/fsck.c
> > > +++ b/fsck/fsck.c
> > > @@ -3299,8 +3299,9 @@ static int chk_and_fix_wp_with_sit(int UNUSED(i), 
> > > void *blkzone, void *opaque)
> > >                                         "segment 0x%x", zone_segno);
> > >                                 ret = f2fs_reset_zone(wpd->dev_index, 
> > > blkz);
> > >                                 if (ret) {
> > > -                                       printf("[FSCK] Write pointer 
> > > reset "
> > > -                                              "failed: %s\n", dev->path);
> > > +                                       ASSERT_MSG("Write pointer reset "
> > > +                                              "failed: %s, err=%d\n",
> > > +                                              dev->path, ret);
> > >                                         return ret;
> > >                                 }
> > >                                 fsck->chk.wp_fixed = 1;
> > > @@ -3326,18 +3327,13 @@ static int chk_and_fix_wp_with_sit(int UNUSED(i), 
> > > void *blkzone, void *opaque)
> > >
> > >         ret = f2fs_finish_zone(wpd->dev_index, blkz);
> > >         if (ret) {
> > > -               u64 fill_sects = blk_zone_length(blkz) -
> > > -                       (blk_zone_wp_sector(blkz) - 
> > > blk_zone_sector(blkz));
> > > -               printf("[FSCK] Finishing zone failed: %s\n", dev->path);
> > > -               ret = dev_fill(NULL, wp_block * F2FS_BLKSIZE,
> > > -                       (fill_sects >> log_sectors_per_block) * 
> > > F2FS_BLKSIZE);
> >
> > This was for zoned devices that don't support the FINISH command. If
> > NULL buf is the problem, how about using zero value filled buffer?
>
> Do we really need to fill them, if it'll be reset later before use?

That's for devices not supporting the FINISH command. Let me make
another patch for it.

>
> >
> > > -               if (ret)
> > > -                       printf("[FSCK] Fill up zone failed: %s\n", 
> > > dev->path);
> > > +               ASSERT_MSG("Finishing zone failed: %s, err=%d\n",
> > > +                       dev->path, ret);
> > > +               return ret;
> > >         }
> > >
> > > -       if (!ret)
> > > -               fsck->chk.wp_fixed = 1;
> > > -       return ret;
> > > +       fsck->chk.wp_fixed = 1;
> > > +       return 0;
> > >  }
> > >
> > >  static void fix_wp_sit_alignment(struct f2fs_sb_info *sbi)
> > > --
> > > 2.46.0.76.ge559c4bf1a-goog
> > >
> > >
> > >
> > > _______________________________________________
> > > Linux-f2fs-devel mailing list
> > > Linux-f2fs-devel@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Reply via email to