0001-libxfs-fix-atomic64_t-detection-on-x86-32-bit-archit.patch

@@ -1,65 +0,0 @@
-From 73ae943b19e658295140f56cb0f0fada2294921d Mon Sep 17 00:00:00 2001
-From: "Darrick J. Wong" <djwong@kernel.org>
-Date: Tue, 12 Sep 2023 12:40:10 -0700
-Subject: [PATCH] libxfs: fix atomic64_t detection on x86 32-bit architectures
-References: bsc#1214416
-
-xfsprogs during compilation tries to detect if liburcu supports atomic
-64-bit ops on the platform it is being compiled on, and if not it falls
-back to using pthread mutex locks.
-
-The detection logic for that fallback relies on _uatomic_link_error()
-which is a link-time trick used by liburcu that will cause compilation
-errors on archs that lack the required support. That only works for the
-generic liburcu code though, and it is not implemented for the
-x86-specific code.
-
-In practice this means that when xfsprogs is compiled on 32-bit x86
-archs will successfully link to liburcu for atomic ops, but liburcu does
-not support atomic64_t on those archs. It indicates this during runtime
-by generating an illegal instruction that aborts execution, and thus
-causes various xfsprogs utils to be segfaulting.
-
-Fix this by executing the liburcu atomic64_t detection code during
-configure instead of only relying on the linker error, so that
-compilation will properly fall back to pthread mutexes on those archs.
-
-Fixes: 7448af588a2e ("libxfs: fix atomic64_t poorly for 32-bit architectures")
-Reported-by: Anthony Iliopoulos <ailiop@suse.com>
-Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
----
- m4/package_urcu.m4 | 9 ++++++++-
- 1 file changed, 8 insertions(+), 1 deletion(-)
-
-diff --git a/m4/package_urcu.m4 b/m4/package_urcu.m4
-index ef116e0cda76..4bb2b886f068 100644
---- a/m4/package_urcu.m4
-+++ b/m4/package_urcu.m4
-@@ -26,7 +26,11 @@ rcu_init();
- #
- # Make sure that calling uatomic_inc on a 64-bit integer doesn't cause a link
- # error on _uatomic_link_error, which is how liburcu signals that it doesn't
--# support atomic operations on 64-bit data types.
-+# support atomic operations on 64-bit data types for its generic
-+# implementation (which relies on compiler builtins). For certain archs
-+# where liburcu carries its own implementation (such as x86_32), it
-+# signals lack of support during runtime by emitting an illegal
-+# instruction, so we also need to check CAA_BITS_PER_LONG to detect that.
- #
- AC_DEFUN([AC_HAVE_LIBURCU_ATOMIC64],
-   [ AC_MSG_CHECKING([for atomic64_t support in liburcu])
-@@ -34,8 +38,11 @@ AC_DEFUN([AC_HAVE_LIBURCU_ATOMIC64],
-     [	AC_LANG_PROGRAM([[
- #define _GNU_SOURCE
- #include <urcu.h>
-+#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)]))
- 	]], [[
- long long f = 3;
-+
-+BUILD_BUG_ON(CAA_BITS_PER_LONG < 64);
- uatomic_inc(&f);
- 	]])
-     ], have_liburcu_atomic64=yes
--- 
-2.42.0
-

0001-repair-shift-inode-back-into-place-if-corrupted-by-b.patch

@@ -1,280 +0,0 @@
-From e2239ef552a48edd33740fec8a005a7ac12dcc80 Mon Sep 17 00:00:00 2001
-From: Jeffrey Mahoney <jeffm@suse.com>
-Date: Tue, 21 Aug 2018 13:41:20 -0400
-Subject: [PATCH] repair: shift inode back into place if corrupted by bad log
- replay
-References: bsc#1105396
-
-SUSE kernels 3.12.74-60.64.40 through 3.12.74-60.64.99 contained
-a regression where xfs_icdinode_t modified di_dmstate to be
-an atomic_t.  Since we only complain if an inode item is too large,
-if a kernel with this patch applied mounted a file system with inode
-items in the log formatted by a kernel without this patch, they would
-be used but would be interpreted using the structure with the atomic_t.
-
-As a result, the inode would be copied incorrectly, corrupting di_dmstate
-and the members that follow it.
-
-For v3 inodes, we can detect that the UUID is shifted forward
-8 bytes and recover di_uuid, di_ino, di_crtime, di_pad2, di_cowextsize,
-di_flags2, and di_lsn.  The UUID and inode number being incorrect
-will trip the verifier on iread, but it will have been flushed from the
-log in a broken state.
-
-di_changecount is lost entirely since half is overwritten by the CRC
-being updated and the other half fell in a hole in the structure.
-di_flags is lost entirely since it is overwritten by the half of
-the generation number.  Half of the generation number is lost since
-it falls in a hole in the structure.
-
-For v2 inodes, the corruption is more limited but impossible to
-detect beyond invalid flags being in use.
-
-Without this fix, xfs_repair will clear the affected inodes, causing
-big problems.
-
-Signed-off-by: Jeff Mahoney <jeffm@suse.com>
----
- repair/dinode.c | 186 ++++++++++++++++++++++++++++++++++++++++++++++--
- 1 file changed, 180 insertions(+), 6 deletions(-)
-
-diff --git a/repair/dinode.c b/repair/dinode.c
-index 8ea919698d14..81238568ac8e 100644
---- a/repair/dinode.c
-+++ b/repair/dinode.c
-@@ -2239,6 +2239,160 @@ _("Bad extent size hint %u on inode %" PRIu64 ", "),
- 	}
- }
- 
-+static int
-+check_shifted_uuid(struct xfs_dinode *dino, xfs_mount_t *mp)
-+{
-+	uint64_t tmp64;
-+	char tmpuuid[16];
-+	uuid_t uuid;
-+
-+	tmp64 = be64_to_cpu(dino->di_ino);
-+	memcpy(tmpuuid, &tmp64, sizeof(tmp64));
-+	memcpy(tmpuuid + 8, &dino->di_uuid, 8);
-+	memcpy(uuid, tmpuuid, 16);
-+
-+	return !platform_uuid_compare(&uuid, &mp->m_sb.sb_meta_uuid);
-+}
-+
-+/*
-+ * There was a kernel that would use incorrectly-formatted log items.
-+ * If it recovered a dirty log, corrupted inodes would result.
-+ * 12 bytes of the inode are completely unrecoverable. Those are
-+ * documented below.
-+ */
-+static void
-+repair_inode_with_bad_atomic(struct xfs_dinode *dino, xfs_mount_t *mp)
-+{
-+	struct xfs_dinode fixed;
-+	struct xfs_legacy_timestamp *lts;
-+	uint64_t tmp64;
-+	uint32_t tmp32;
-+	char tmpuuid[16];
-+	char *tmpptr;
-+
-+	uuid_t uuid;
-+
-+	tmp64 = be64_to_cpu(dino->di_ino);
-+	memcpy(tmpuuid, &tmp64, sizeof(tmp64));
-+	tmpptr = (char *)dino->di_uuid;
-+	memcpy(tmpuuid + 8, tmpptr, 8);
-+	memcpy(uuid, tmpuuid, 16);
-+
-+	memcpy(&fixed, dino, sizeof(fixed));
-+	memcpy(&fixed.di_uuid, uuid, sizeof(uuid));
-+
-+	tmp32 = *(uint32_t *)&dino->di_pad2[4];
-+	lts = (struct xfs_legacy_timestamp *)&(fixed.di_crtime);
-+	lts->t_sec = cpu_to_be32(tmp32);
-+	tmp32 = *(uint32_t *)&dino->di_pad2[8];
-+	lts->t_nsec = cpu_to_be32(tmp32);
-+
-+	tmp64 = be32_to_cpu(((struct xfs_legacy_timestamp *)(&(dino->di_crtime)))->t_nsec);
-+	tmp64 <<= 32;
-+	tmp64 |= be32_to_cpu(((struct xfs_legacy_timestamp *)(&(dino->di_crtime)))->t_sec);
-+	fixed.di_ino = cpu_to_be64(tmp64);
-+
-+	tmp64 = be64_to_cpu(fixed.di_ino);
-+
-+	memcpy(fixed.di_pad2 + 8, dino->di_pad2, 4);
-+
-+	tmp32 = be32_to_cpu(dino->di_cowextsize);
-+	memcpy(fixed.di_pad2 + 4, &tmp32, 4);
-+
-+	tmp64 = be64_to_cpu(dino->di_flags2);
-+	tmp32 = tmp64 >> 32;
-+	memcpy(fixed.di_pad2, &tmp32, 4);
-+
-+	fixed.di_cowextsize = cpu_to_be32(tmp64);
-+	fixed.di_flags2 = dino->di_lsn;
-+	fixed.di_lsn = dino->di_changecount;
-+
-+	/*
-+	 * This is lost entirely.  Half falls in padding and half
-+	 * is overwritten by the CRC.
-+	*/
-+	fixed.di_changecount = 0;
-+
-+#if __BYTE_ORDER == __LITTLE_ENDIAN
-+	/*
-+	 * Half of the generation number is lost, but it's the high bits.
-+	 * Pick a high number and hope for the best.
-+	 */
-+	tmp32 = 0xff000000;
-+	tmp32 |= be16_to_cpu(dino->di_flags);
-+	fixed.di_gen = cpu_to_be32(tmp32);
-+#else
-+	/*
-+	 * Half of the generation number is lost, but it's the low bits,
-+	 * so we can fake it.
-+	 */
-+	tmp32 = be16_to_cpu(dino->di_flags) + 1;
-+	tmp32 <<= 16;
-+	fixed.di_gen = cpu_to_be32(tmp32);
-+#endif
-+
-+	/*
-+	 * The flags are lost since the atomic_t was 32-bit and we
-+	 * only keep 16.
-+	 */
-+	fixed.di_flags = 0;
-+
-+	memcpy(dino, &fixed, sizeof(*dino));
-+	xfs_dinode_calc_crc(mp, dino);
-+}
-+
-+static int
-+process_dinode_int(xfs_mount_t *mp, struct xfs_dinode *dino, xfs_agnumber_t agno,
-+		   xfs_agino_t ino, int was_free, int *dirty, int *used,
-+		   int verify_mode, int uncertain, int ino_discovery,
-+		   int check_dups, int extra_attr_check, int *isa_dir,
-+		   xfs_ino_t *parent, int recurse);
-+
-+static int
-+handle_malformed_inode(xfs_mount_t *mp, struct xfs_dinode *dino,
-+		       xfs_agnumber_t agno, xfs_agino_t ino, int was_free,
-+		       int *dirty, int *used, int verify_mode, int uncertain,
-+		       int ino_discovery, int check_dups, int extra_attr_check,
-+		       int *isa_dir, xfs_ino_t *parent)
-+{
-+	struct xfs_dinode save;
-+	int retval;
-+	xfs_ino_t lino = XFS_AGINO_TO_INO(mp, agno, ino);
-+
-+	if (!uncertain)
-+		do_warn(_("malformed inode %" PRIu64 " found%c"),
-+			lino, verify_mode ? '\n' : ',');
-+
-+	/*
-+	 * We can't just pass a local copy to verify since we need the
-+	 * data fork to check directories.
-+	 */
-+	if (verify_mode || no_modify)
-+		memcpy(&save, dino, sizeof(*dino));
-+
-+	repair_inode_with_bad_atomic(dino, mp);
-+	retval = process_dinode_int(mp, dino, agno, ino, was_free, dirty,
-+				    used, verify_mode, uncertain, ino_discovery,
-+				    check_dups, extra_attr_check,
-+				    isa_dir, parent, 1);
-+
-+	if (verify_mode || no_modify) {
-+		memcpy(dino, &save, sizeof(*dino));
-+		*dirty = 0;
-+	}
-+
-+	if (retval == 0 && !verify_mode) {
-+		if (no_modify)
-+			do_warn(_(" would repair\n"));
-+		else {
-+			do_warn(_(" repairing\n"));
-+			*dirty = 1;
-+		}
-+	}
-+
-+	return retval;
-+}
-+
- /*
-  * returns 0 if the inode is ok, 1 if the inode is corrupt
-  * check_dups can be set to 1 *only* when called by the
-@@ -2263,7 +2417,8 @@ process_dinode_int(xfs_mount_t *mp,
- 					 * duplicate blocks		*/
- 		int extra_attr_check, /* 1 == do attribute format and value checks */
- 		int *isa_dir,		/* out == 1 if inode is a directory */
--		xfs_ino_t *parent)	/* out -- parent if ino is a dir */
-+		xfs_ino_t *parent,	/* out -- parent if ino is a dir */
-+		int recurse)
- {
- 	xfs_rfsblock_t		totblocks = 0;
- 	xfs_rfsblock_t		atotblocks = 0;
-@@ -2379,6 +2534,25 @@ process_dinode_int(xfs_mount_t *mp,
- 	 * memory and hence invalidated the CRC.
- 	 */
- 	if (xfs_has_crc(mp)) {
-+		int good_uuid = 1;
-+
-+		if (platform_uuid_compare(&dino->di_uuid,
-+					  &mp->m_sb.sb_meta_uuid))
-+			good_uuid = 0;
-+
-+		/*
-+		 * Only check to see if it's a malformed inode if it has
-+		 * a valid magic, crc, and version and an invalid uuid.
-+		 */
-+		if (!good_uuid && !retval && !recurse &&
-+		    check_shifted_uuid(dino, mp))
-+			return handle_malformed_inode(mp, dino, agno, ino,
-+						was_free, dirty, used,
-+						verify_mode, uncertain,
-+						ino_discovery, check_dups,
-+						extra_attr_check,
-+						isa_dir, parent);
-+
- 		if (be64_to_cpu(dino->di_ino) != lino) {
- 			if (!uncertain)
- 				do_warn(
-@@ -2389,8 +2563,7 @@ _("inode identifier %llu mismatch on inode %" PRIu64 "\n"),
- 				return 1;
- 			goto clear_bad_out;
- 		}
--		if (platform_uuid_compare(&dino->di_uuid,
--					  &mp->m_sb.sb_meta_uuid)) {
-+		if (!good_uuid) {
- 			if (!uncertain)
- 				do_warn(
- 			_("UUID mismatch on inode %" PRIu64 "\n"), lino);
-@@ -2952,7 +3125,8 @@ process_dinode(
- #endif
- 	return process_dinode_int(mp, dino, agno, ino, was_free, dirty, used,
- 				verify_mode, uncertain, ino_discovery,
--				check_dups, extra_attr_check, isa_dir, parent);
-+				check_dups, extra_attr_check, isa_dir, parent,
-+				0);
- }
- 
- /*
-@@ -2979,7 +3153,7 @@ verify_dinode(
- 
- 	return process_dinode_int(mp, dino, agno, ino, 0, &dirty, &used,
- 				verify_mode, uncertain, ino_discovery,
--				check_dups, 0, &isa_dir, &parent);
-+				check_dups, 0, &isa_dir, &parent, 0);
- }
- 
- /*
-@@ -3005,5 +3179,5 @@ verify_uncertain_dinode(
- 
- 	return process_dinode_int(mp, dino, agno, ino, 0, &dirty, &used,
- 				verify_mode, uncertain, ino_discovery,
--				check_dups, 0, &isa_dir, &parent);
-+				check_dups, 0, &isa_dir, &parent, 0);
- }
--- 
-2.36.1
-

xfsprogs-6.4.0.tar.sign

@@ -1,7 +0,0 @@
------BEGIN PGP SIGNATURE-----
-
-iIUEABYIAC0WIQT6QG4gav94c4l8aGS0VhjDaiT9IwUCZLfVMw8cY2VtQGtlcm5l
-bC5vcmcACgkQtFYYw2ok/SNKPwEA6fsxp1TRbPXQn6a605fU3cE6WjcqDLej3zYa
-lx91BnABAKIWP14Rd7KPbCH4ezAPydFZxSn26trKqPxzAFOZ/skP
-=mkU4
------END PGP SIGNATURE-----

xfsprogs-6.5.0.tar.sign

@@ -0,0 +1,7 @@
+-----BEGIN PGP SIGNATURE-----
+
+iIUEABYIAC0WIQT6QG4gav94c4l8aGS0VhjDaiT9IwUCZSfZXQ8cY2VtQGtlcm5l
+bC5vcmcACgkQtFYYw2ok/SOFogD+MZ4LooqWPZ9aNEfw/GEtiJJzlVMwYHFcgqDK
+xgilyusBAKOZeJc4SiFarTv2PnIRuMDxhwaFm0Sy+ngiFn0whlED
+=k6/7
+-----END PGP SIGNATURE-----

xfsprogs.changes

@@ -1,3 +1,30 @@
+-------------------------------------------------------------------
+Fri Oct 13 14:45:05 UTC 2023 - Anthony Iliopoulos <ailiop@suse.com>
+
+- update to 6.5.0
+  - libxfs: fix atomic64_t detection on x86_32
+  - libxfs: use XFS_IGET_CREATE when creating new files
+  - libfrog: fix overly sleep workqueues
+  - xfs_db: use directio for device access
+  - libxfs: make platform_set_blocksize optional with directio
+  - mkfs: add a config file for 6.6 LTS kernels
+  - mkfs: enable reverse mapping by default
+  - mkfs: enable large extent counts by default
+  - xfs_db: create unlinked inodes
+  - xfs_db: dump unlinked buckets
+  - xfsprogs: don't allow udisks to automount XFS filesystems with no prompt
+  - xfs_repair: fix repair failure caused by dirty flag being abnormally set on buffer
+
+- drop 0001-repair-shift-inode-back-into-place-if-corrupted-by-b.patch
+
+  This is now upstream.
+
+- drop 0001-repair-shift-inode-back-into-place-if-corrupted-by-b.patch
+
+  This was a fix for a regression that occurred in SLE12 SP1. We can
+  safely drop this, as upgrading to SLE15 (and later) is only supported
+  from SLE12-SP4.
+
 -------------------------------------------------------------------
 Fri Sep 22 12:56:31 UTC 2023 - Anthony Iliopoulos <ailiop@suse.com>
 

xfsprogs.spec

@@ -25,7 +25,7 @@
 %endif
 %define libname libhandle1
 Name:           xfsprogs
-Version:        6.4.0
+Version:        6.5.0
 Release:        0
 Summary:        Utilities for managing the XFS file system
 License:        GPL-2.0-or-later
@@ -37,8 +37,6 @@ Source2:        %{name}.keyring
 Source3:        module-setup.sh.in
 Source4:        dracut-fsck-help.txt
 Patch0:         xfsprogs-docdir.diff
-Patch1:         0001-repair-shift-inode-back-into-place-if-corrupted-by-b.patch
-Patch2:         0001-libxfs-fix-atomic64_t-detection-on-x86-32-bit-archit.patch
 BuildRequires:  autoconf
 BuildRequires:  automake
 BuildRequires:  libblkid-devel
@@ -111,8 +109,6 @@ on xfs filesystems.
 %prep
 %setup -q
 %patch0 -p1
-%patch1 -p1
-%patch2 -p1
 
 %build
 aclocal -I m4