[PATCHv2 8/10] xfs: avoid repeated pointer dereferences

[PATCHv2 8/10] xfs: avoid repeated pointer dereferences

[Alex Elder]

In xlog_find_cycle_start() use a local variable for some repeated
operations rather than constantly updating the memory location
whose address is passed in.

Signed-off-by: Alex Elder <aelder@sgi.com>

---
fs/xfs/xfs_log_recover.c |   83 ++++++++++++++++++++++++-----------------------
 fs/xfs/xfs_log_recover.c |   26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

Index: b/fs/xfs/xfs_log_recover.c
===================================================================
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -354,26 +354,28 @@ xlog_find_cycle_start(
 {
 	xfs_caddr_t	offset;
 	xfs_daddr_t	mid_blk;
+	xfs_daddr_t	end_blk;
 	uint		mid_cycle;
 	int		error;
 
-	mid_blk = BLK_AVG(first_blk, *last_blk);
-	while (mid_blk != first_blk && mid_blk != *last_blk) {
+	ASSERT(last_blk != NULL);
+	end_blk = *last_blk;
+	mid_blk = BLK_AVG(first_blk, end_blk);
+	while (mid_blk != first_blk && mid_blk != end_blk) {
 		error = xlog_bread(log, mid_blk, 1, bp, &offset);
 		if (error)
 			return error;
 		mid_cycle = xlog_get_cycle(offset);
-		if (mid_cycle == cycle) {
-			*last_blk = mid_blk;
-			/* last_half_cycle == mid_cycle */
-		} else {
-			first_blk = mid_blk;
-			/* first_half_cycle == mid_cycle */
-		}
-		mid_blk = BLK_AVG(first_blk, *last_blk);
+		if (mid_cycle == cycle)
+			end_blk = mid_blk;   /* last_half_cycle == mid_cycle */
+		else
+			first_blk = mid_blk; /* first_half_cycle == mid_cycle */
+		mid_blk = BLK_AVG(first_blk, end_blk);
 	}
-	ASSERT((mid_blk == first_blk && mid_blk+1 == *last_blk) ||
-	       (mid_blk == *last_blk && mid_blk-1 == first_blk));
+	ASSERT((mid_blk == first_blk && mid_blk+1 == end_blk) ||
+	       (mid_blk == end_blk && mid_blk-1 == first_blk));
+
+	*last_blk = end_blk;
 
 	return 0;
 }


_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCHv2 8/10] xfs: avoid repeated pointer dereferences

[Dave Chinner]

On Fri, Apr 09, 2010 at 05:30:42PM -0500, Alex Elder wrote:
> In xlog_find_cycle_start() use a local variable for some repeated
> operations rather than constantly updating the memory location
> whose address is passed in.

Won't the compiler optimise that out for us? i.e. does the dissassembly
of the function look any better before and after this change?

> Signed-off-by: Alex Elder <aelder@sgi.com>
> 
> ---
> fs/xfs/xfs_log_recover.c |   83 ++++++++++++++++++++++++-----------------------
>  fs/xfs/xfs_log_recover.c |   26 ++++++++++++++------------
>  1 file changed, 14 insertions(+), 12 deletions(-)
> 
> Index: b/fs/xfs/xfs_log_recover.c
> ===================================================================
> --- a/fs/xfs/xfs_log_recover.c
> +++ b/fs/xfs/xfs_log_recover.c
> @@ -354,26 +354,28 @@ xlog_find_cycle_start(
>  {
>  	xfs_caddr_t	offset;
>  	xfs_daddr_t	mid_blk;
> +	xfs_daddr_t	end_blk;
>  	uint		mid_cycle;
>  	int		error;
>  
> -	mid_blk = BLK_AVG(first_blk, *last_blk);
> -	while (mid_blk != first_blk && mid_blk != *last_blk) {
> +	ASSERT(last_blk != NULL);
> +	end_blk = *last_blk;

FWIW, there is no need for that ASSERT there - the machine will
panic on the very next line, anyway....

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCHv2 8/10] xfs: avoid repeated pointer dereferences

[Alex Elder]

On Mon, 2010-04-12 at 16:56 +1000, Dave Chinner wrote:
> On Fri, Apr 09, 2010 at 05:30:42PM -0500, Alex Elder wrote:
> > In xlog_find_cycle_start() use a local variable for some repeated
> > operations rather than constantly updating the memory location
> > whose address is passed in.
> 
> Won't the compiler optimise that out for us? i.e. does the dissassembly
> of the function look any better before and after this change?

I doubt it *can* optimize that.  Doing so would change
the way the function interacts semantically with the
pointed-to memory.  Still, I performed a quick check
to be sure (on an x86_64) and the compiled code really
does de-reference the pointer for both reads from and
writes to memory.

> > Signed-off-by: Alex Elder <aelder@sgi.com>
> > 
> > ---
> > fs/xfs/xfs_log_recover.c |   83 ++++++++++++++++++++++++-----------------------
> >  fs/xfs/xfs_log_recover.c |   26 ++++++++++++++------------
> >  1 file changed, 14 insertions(+), 12 deletions(-)
> > 
> > Index: b/fs/xfs/xfs_log_recover.c
> > ===================================================================
> > --- a/fs/xfs/xfs_log_recover.c
> > +++ b/fs/xfs/xfs_log_recover.c
> > @@ -354,26 +354,28 @@ xlog_find_cycle_start(
> >  {
> >  	xfs_caddr_t	offset;
> >  	xfs_daddr_t	mid_blk;
> > +	xfs_daddr_t	end_blk;
> >  	uint		mid_cycle;
> >  	int		error;
> >  
> > -	mid_blk = BLK_AVG(first_blk, *last_blk);
> > -	while (mid_blk != first_blk && mid_blk != *last_blk) {
> > +	ASSERT(last_blk != NULL);
> > +	end_blk = *last_blk;
> 
> FWIW, there is no need for that ASSERT there - the machine will
> panic on the very next line, anyway....

Agreed.  I make a habit of stating assumptions about
passed-in arguments via assertions.  Not a big deal,
I'll take it out.

Still soliciting a reviewed-by on this one.

					-Alex


_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs