Attachment #181514 for bug #218174

Lines 3-14 Link Here

(-)Makefile (-2 / +5 lines)
3		3
4	PORTNAME= ffs2recov	4	PORTNAME= ffs2recov
5	PORTVERSION= 1.0	5	PORTVERSION= 1.0
		6	PORTREVISION= 1
6	CATEGORIES= sysutils	7	CATEGORIES= sysutils
7	MASTER_SITES= http://www.exit.com/Archives/FreeBSD/	8	MASTER_SITES= http://BSDforge.com/projects/source/sysutils/ffs2recov/
8		9
9	MAINTAINER= ports@FreeBSD.org	10	MAINTAINER= portmaster@bsdforge.com
10	COMMENT= Utility to recover UFS2 filesystems	11	COMMENT= Utility to recover UFS2 filesystems
11		12
		13	LICENSE= BSD2CLAUSE
		14
12	USES= tar:bzip2 uidfix	15	USES= tar:bzip2 uidfix
13	WRKSRC= ${WRKDIR}/${PORTNAME}	16	WRKSRC= ${WRKDIR}/${PORTNAME}
14		17




--- inode.c.orig	2017-04-05 19:51:41 UTC
+++ inode.c
@@ -52,6 +52,10 @@
 #define NO_IN_EXTERNS
 #include "ffs2recov.h"
 
+#if __FreeBSD_version >= 1200021
+#define NDADDR UFS_NDADDR
+#endif
+
 static uint64_t direct_maxblk;		/* Max block # for a direct block.    */
 static uint64_t single_maxblk;		/* Max block # for a single indirect. */
 static uint64_t double_maxblk;		/* Max block # for a double indirect. */




This is the UFS2 version of ffsrecov, heavily (and I do mean _heavily_)
based on John-Mark Gurney's program of the same name. It does basically
the same thing, only it's a little more resistant to crashes caused by
bad pointers, offsets and the like, and it does a little more than his
did. Don't contact him for problems with this program, it's definitely
_my_ fault if it breaks.

This program is not ready for prime time. It has some shortfalls, it
has a bunch of new options that are mostly undocumented and the manpage
could stand to be rewritten. One _good_ thing is that it now uses the
libufs library and is therefore not as dependent on carrying around
low-level code.

On the other hand, it worked for me.  Using this tool, I was able to
recover almost all of a several-hundred-gigabyte file system that had
been stomped by a misconfigured RAID controller. (That's why I wrote the
thing in the first place, in fact.)  With the right knowledge and a lot
of patience, it is possible to recover most or all of a trashed file
system, at least if it's not _too_ trashed.

I'm releasing it under the Berkeley two-clause license in the hope that
someone with more time will pick it up, polish it and make something a
little more useful out of it.

Frank Mayhar
frank@exit.com

WWW: http://BSDforge.com/projects/sysutils/ffs2recov/

Return to bug 218174

Line 0 Link Here

(-)files/patch-inode.c (+13 lines)
	1	--- inode.c.orig 2017-04-05 19:51:41 UTC
	2	+++ inode.c
	3	@@ -52,6 +52,10 @@
	4	#define NO_IN_EXTERNS
	5	#include "ffs2recov.h"
	6
	7	+#if __FreeBSD_version >= 1200021
	8	+#define NDADDR UFS_NDADDR
	9	+#endif
	10	+
	11	static uint64_t direct_maxblk; /* Max block # for a direct block. */
	12	static uint64_t single_maxblk; /* Max block # for a single indirect. */
	13	static uint64_t double_maxblk; /* Max block # for a double indirect. */

Lines 1-24 Link Here

(-)pkg-descr (-17 / +21 lines)
1	This is the UFS2 version of ffsrecov, heavily (and I do mean _heavily_) based	1	This is the UFS2 version of ffsrecov, heavily (and I do mean _heavily_)
2	on John-Mark Gurney's program of the same name. It does basically the same	2	based on John-Mark Gurney's program of the same name. It does basically
3	thing, only it's a little more resistant to crashes caused by bad pointers,	3	the same thing, only it's a little more resistant to crashes caused by
4	offsets and the like, and it does a little more than his did. Don't contact	4	bad pointers, offsets and the like, and it does a little more than his
5	him for problems with this program, it's definitely _my_ fault if it breaks.	5	did. Don't contact him for problems with this program, it's definitely
		6	_my_ fault if it breaks.
6		7
7	This program is not ready for prime time. It has some shortfalls, it has a	8	This program is not ready for prime time. It has some shortfalls, it
8	bunch of new options that are mostly undocumented and the manpage could	9	has a bunch of new options that are mostly undocumented and the manpage
9	stand to be rewritten. One _good_ thing is that it now uses the libufs	10	could stand to be rewritten. One _good_ thing is that it now uses the
10	library and is therefore not as dependent on carrying around low-level code.	11	libufs library and is therefore not as dependent on carrying around
		12	low-level code.
11		13
12	On the other hand, it worked for me. Using this tool, I was able to recover	14	On the other hand, it worked for me. Using this tool, I was able to
13	almost all of a several-hundred-gigabyte file system that had been stomped	15	recover almost all of a several-hundred-gigabyte file system that had
14	by a misconfigured RAID controller. (That's why I wrote the thing in the	16	been stomped by a misconfigured RAID controller. (That's why I wrote the
15	first place, in fact.) With the right knowledge and a lot of patience,	17	thing in the first place, in fact.) With the right knowledge and a lot
16	it is possible to recover most or all of a trashed file system, at least if	18	of patience, it is possible to recover most or all of a trashed file
17	it's not _too_ trashed.	19	system, at least if it's not _too_ trashed.
18		20
19	I'm releasing it under the Berkeley two-clause license in the hope that	21	I'm releasing it under the Berkeley two-clause license in the hope that
20	someone with more time will pick it up, polish it and make something	22	someone with more time will pick it up, polish it and make something a
21	a little more useful out of it.	23	little more useful out of it.
22		24
23	Frank Mayhar	25	Frank Mayhar
24	frank@exit.com	26	frank@exit.com
		27
		28	WWW: http://BSDforge.com/projects/sysutils/ffs2recov/