[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20181130203611.GC71781@jaegeuk-macbookpro.roam.corp.google.com>
Date: Fri, 30 Nov 2018 12:36:11 -0800
From: Jaegeuk Kim <jaegeuk@...nel.org>
To: Chao Yu <yuchao0@...wei.com>
Cc: linux-kernel@...r.kernel.org,
linux-f2fs-devel@...ts.sourceforge.net
Subject: Re: [f2fs-dev] [PATCH] f2fs: add an ioctl() to explicitly trigger
fsck later
On 11/30, Chao Yu wrote:
> On 2018/11/29 9:52, Jaegeuk Kim wrote:
> > This adds an option in ioctl(F2FS_IOC_SHUTDOWN) in order to trigger fsck by
> > setting a NEED_FSCK flag.
> >
> > Signed-off-by: Jaegeuk Kim <jaegeuk@...nel.org>
> > ---
> > fs/f2fs/f2fs.h | 1 +
> > fs/f2fs/file.c | 7 +++++++
> > 2 files changed, 8 insertions(+)
> >
> > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > index aa500239baf2..7cec897146a3 100644
> > --- a/fs/f2fs/f2fs.h
> > +++ b/fs/f2fs/f2fs.h
> > @@ -417,6 +417,7 @@ static inline bool __has_cursum_space(struct f2fs_journal *journal,
> > #define F2FS_GOING_DOWN_METASYNC 0x1 /* going down with metadata */
> > #define F2FS_GOING_DOWN_NOSYNC 0x2 /* going down */
> > #define F2FS_GOING_DOWN_METAFLUSH 0x3 /* going down with meta flush */
> > +#define F2FS_GOING_DOWN_NEED_FSCK 0x4 /* going down to trigger fsck */
>
> Why not add a new ioctl interface for this? F2FS_GOING_DOWN_ prefix implies
> filesystem will shutdown, IMO, we'd better to keep all sub-interfaces being
> consistent in f2fs_ioc_shutdown().
I'm thinking to use this for QA as device shutdown tests.
>
> >
> > #if defined(__KERNEL__) && defined(CONFIG_COMPAT)
> > /*
> > diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> > index ff82350a2c55..ca9bdbb8651b 100644
> > --- a/fs/f2fs/file.c
> > +++ b/fs/f2fs/file.c
> > @@ -1966,6 +1966,13 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg)
> > f2fs_stop_checkpoint(sbi, false);
> > set_sbi_flag(sbi, SBI_IS_SHUTDOWN);
> > break;
> > + case F2FS_GOING_DOWN_NEED_FSCK:
> > + set_sbi_flag(sbi, SBI_NEED_FSCK);
> > + /* do checkpoint only */
> > + ret = f2fs_sync_fs(sb, 1);
> > + if (ret)
> > + goto out;
>
> In large-sized image, it may take long time to trigger full scan during
> boot, so I'd like to ask how often we set this flag?
Based on the use of shutdown ioctl, I'll use this for testing purpose, so it
won't be used for regular cases.
>
> Thanks,
>
> > + break;
> > default:
> > ret = -EINVAL;
> > goto out;
> >
Powered by blists - more mailing lists