Re: [PATCH] xfs: clean up calculation of LR header blocks

[Gao Xiang <hsiangkao@redhat.com>]

Hi Brian,

On Wed, Sep 02, 2020 at 11:26:56AM -0400, Brian Foster wrote:
> On Wed, Sep 02, 2020 at 10:09:23PM +0800, Gao Xiang wrote:
> > Let's use DIV_ROUND_UP() to calculate log record header
> > blocks as what did in xlog_get_iclog_buffer_size() and
> > also wrap up a common helper for log recovery code.
> > 
> > Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
> > ---
> >  fs/xfs/libxfs/xfs_log_format.h | 10 +++++++++
> >  fs/xfs/xfs_log.c               |  4 +---
> >  fs/xfs/xfs_log_recover.c       | 37 ++++++++--------------------------
> >  3 files changed, 19 insertions(+), 32 deletions(-)
> > 
> ...
> > diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
> > index e2ec91b2d0f4..5fecc0c7aeb2 100644
> > --- a/fs/xfs/xfs_log_recover.c
> > +++ b/fs/xfs/xfs_log_recover.c
> > @@ -489,15 +489,10 @@ xlog_find_verify_log_record(
> >  	 * reset last_blk.  Only when last_blk points in the middle of a log
> >  	 * record do we update last_blk.
> >  	 */
> > -	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) {
> > -		uint	h_size = be32_to_cpu(head->h_size);
> > -
> > -		xhdrs = h_size / XLOG_HEADER_CYCLE_SIZE;
> > -		if (h_size % XLOG_HEADER_CYCLE_SIZE)
> > -			xhdrs++;
> > -	} else {
> > +	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb))
> > +		xhdrs = xlog_logv2_rec_hblks(head);
> > +	else
> >  		xhdrs = 1;
> > -	}
> 
> Can we fold the _haslogv2() check into the helper as well (then perhaps
> drop the 'v2' from the name)? Otherwise looks like a nice cleanup to
> me..

Thanks for your suggestion.

I would have done the same at the beginning, but the case
in xlog_do_recovery_pass() from the current codebase is a
bit complex for now.

I will try to simplify the logic in xlog_do_recovery_pass()
tomorrow, but if it makes more complex.. I'd suggest leave
it as-is...

Thanks,
Gao Xiang

> 
> Brian
> 
> >  
> >  	if (*last_blk - i + extra_bblks !=
> >  	    BTOBB(be32_to_cpu(head->h_len)) + xhdrs)
> > @@ -1184,21 +1179,10 @@ xlog_check_unmount_rec(
> >  	 * below. We won't want to clear the unmount record if there is one, so
> >  	 * we pass the lsn of the unmount record rather than the block after it.
> >  	 */
> > -	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) {
> > -		int	h_size = be32_to_cpu(rhead->h_size);
> > -		int	h_version = be32_to_cpu(rhead->h_version);
> > -
> > -		if ((h_version & XLOG_VERSION_2) &&
> > -		    (h_size > XLOG_HEADER_CYCLE_SIZE)) {
> > -			hblks = h_size / XLOG_HEADER_CYCLE_SIZE;
> > -			if (h_size % XLOG_HEADER_CYCLE_SIZE)
> > -				hblks++;
> > -		} else {
> > -			hblks = 1;
> > -		}
> > -	} else {
> > +	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb))
> > +		hblks = xlog_logv2_rec_hblks(rhead);
> > +	else
> >  		hblks = 1;
> > -	}
> >  
> >  	after_umount_blk = xlog_wrap_logbno(log,
> >  			rhead_blk + hblks + BTOBB(be32_to_cpu(rhead->h_len)));
> > @@ -3016,15 +3000,10 @@ xlog_do_recovery_pass(
> >  			}
> >  		}
> >  
> > -		if ((be32_to_cpu(rhead->h_version) & XLOG_VERSION_2) &&
> > -		    (h_size > XLOG_HEADER_CYCLE_SIZE)) {
> > -			hblks = h_size / XLOG_HEADER_CYCLE_SIZE;
> > -			if (h_size % XLOG_HEADER_CYCLE_SIZE)
> > -				hblks++;
> > +		hblks = xlog_logv2_rec_hblks(rhead);
> > +		if (hblks != 1) {
> >  			kmem_free(hbp);
> >  			hbp = xlog_alloc_buffer(log, hblks);
> > -		} else {
> > -			hblks = 1;
> >  		}
> >  	} else {
> >  		ASSERT(log->l_sectBBsize == 1);
> > -- 
> > 2.18.1
> > 
>