linux-ext4.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] E2fsprogs/filefrag: fix filefrag regression
@ 2012-06-14  8:44 Liu Bo
  2012-07-17 21:57 ` Eric Sandeen
  2012-07-28 21:34 ` Theodore Ts'o
  0 siblings, 2 replies; 3+ messages in thread
From: Liu Bo @ 2012-06-14  8:44 UTC (permalink / raw)
  To: linux-ext4

filefrag has several bugs:

1.
$ touch f1
$ filefrag f1
f1: 1 extent found  ----> bug!
$ filefrag -v f1
Filesystem type is: ef53
File size of f1 is 0 (0 blocks, blocksize 4096)
f1: 0 extents found

2.
$ truncate -s 1m f2
$ filefrag f2
f2: 1 extent found  ----> bug!
$ filefrag -v f2
Filesystem type is: ef53
File size of f2 is 1048576 (256 blocks, blocksize 4096)
f2: 0 extents found

3.
$ for i in `seq 11 -2 0`; do dd if=/dev/zero of=f4 bs=4k count=1 seek=$i conv=notrunc oflag=sync &>/dev/null; done
$ ll f4
-rw-r--r-- 1 root root 49152 Jun  9 15:09 f4
$ filefrag f4
f4: 7 extents found
$ filefrag -v f4
Filesystem type is: ef53
File size of f4 is 49152 (12 blocks, blocksize 4096)
 ext logical physical expected length flags
   0       1  1109993               1
   1       3  1109992  1109994      1
   2       5  1109991  1109993      1
   3       7  1109990  1109992      1
   4       9  1109989  1109991      1
   5      11  1108207  1109990      1 eof
f4: 7 extents found  --------------------------> but we only have 6 extents, bug!

All of these bugs come from the fact that we've made a mistake on calculating
total extents:
o   we set 1 as default for 'total extents', and this will report 1 extent found
    even when we don't get any extent from fiemap.
o   if our first extent does not start from 0(logical addr), total extents will
    be one more than what it should be.

Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
---
 misc/filefrag.c |   25 ++++++++++---------------
 1 files changed, 10 insertions(+), 15 deletions(-)

diff --git a/misc/filefrag.c b/misc/filefrag.c
index 0583e91..af008fb 100644
--- a/misc/filefrag.c
+++ b/misc/filefrag.c
@@ -175,7 +175,7 @@ static int filefrag_fiemap(int fd, int blk_shift, int *num_extents)
 	unsigned int i;
 	static int fiemap_incompat_printed;
 	int fiemap_header_printed = 0;
-	int tot_extents = 1, n = 0;
+	int tot_extents = 0, n = 0;
 	int last = 0;
 	int rc;
 
@@ -201,25 +201,17 @@ static int filefrag_fiemap(int fd, int blk_shift, int *num_extents)
 			return rc;
 		}
 
+		/* If 0 extents are returned, then more ioctls are not needed */
+		if (fiemap->fm_mapped_extents == 0)
+			break;
+
 		if (verbose && !fiemap_header_printed) {
-			/*
-			 * No extents on first call?
-			 * Skip header and show 0 extents.
-			 */
-			if (fiemap->fm_mapped_extents == 0) {
-				*num_extents = 0;
-				goto out;
-			}
 			printf(" ext %*s %*s %*s length flags\n", logical_width,
 			       "logical", physical_width, "physical",
 			       physical_width, "expected");
 			fiemap_header_printed = 1;
 		}
 
-		/* If 0 extents are returned, then more ioctls are not needed */
-		if (fiemap->fm_mapped_extents == 0)
-			break;

^ permalink raw reply related	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-07-28 21:34 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-06-14  8:44 [PATCH] E2fsprogs/filefrag: fix filefrag regression Liu Bo
2012-07-17 21:57 ` Eric Sandeen
2012-07-28 21:34 ` Theodore Ts'o

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).