[pve-devel] [PATCH zfsonlinux 0/2] update to ZFS 2.1.3

[pve-devel] [PATCH zfsonlinux 0/2] update to ZFS 2.1.3

[Stoiko Ivanov]

updates the submodule to the recently release zfs-2.1.3

* changeset does not look too exciting (kernel-compatibility fixes, testsuite,
  FreeBSD)
* the addition of the systemd timer+service for regular scrubbing is something
  user-visible (but already present in upstream and our debian packages for a
  while)

* the zvol_open lock inversion we cherry-picked is contained in the new tag
* zed default for mailing has changed upstream:
  e56dffe4b5cd00df2d46b75a80d3d17c8fe70eec
* the changes to the other patches were due to some refactorings:
  745a7f78da95d1969613cd55fcb33df1b46c1e66

tested with the latest 5.13 kernel on 2 VMs (running ztest)

Stoiko Ivanov (2):
  update submodule and patches to ZFS 2.1.3
  add zfs-scrub timers+services to zfsutils-linux

 ...META-and-DCH-consistency-in-autoconf.patch |   4 +-
 debian/patches/0005-Enable-zed-emails.patch   |  17 +-
 ...md-unit-for-importing-specific-pools.patch |   2 +-
 .../0012-Fix-zvol_open-lock-inversion.patch   | 212 ------------------
 debian/patches/series                         |   1 -
 debian/zfsutils-linux.install                 |   3 +
 upstream                                      |   2 +-
 7 files changed, 11 insertions(+), 230 deletions(-)
 delete mode 100644 debian/patches/0012-Fix-zvol_open-lock-inversion.patch

-- 
2.30.2

[pve-devel] [PATCH zfsonlinux 1/2] update submodule and patches to ZFS 2.1.3

[Stoiko Ivanov]

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
---
 ...META-and-DCH-consistency-in-autoconf.patch |   4 +-
 debian/patches/0005-Enable-zed-emails.patch   |  17 +-
 ...md-unit-for-importing-specific-pools.patch |   2 +-
 .../0012-Fix-zvol_open-lock-inversion.patch   | 212 ------------------
 debian/patches/series                         |   1 -
 upstream                                      |   2 +-
 6 files changed, 8 insertions(+), 230 deletions(-)
 delete mode 100644 debian/patches/0012-Fix-zvol_open-lock-inversion.patch

diff --git a/debian/patches/0001-Check-for-META-and-DCH-consistency-in-autoconf.patch b/debian/patches/0001-Check-for-META-and-DCH-consistency-in-autoconf.patch
index c1f6a4bb..504bb998 100644
--- a/debian/patches/0001-Check-for-META-and-DCH-consistency-in-autoconf.patch
+++ b/debian/patches/0001-Check-for-META-and-DCH-consistency-in-autoconf.patch
@@ -10,7 +10,7 @@ Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
  1 file changed, 29 insertions(+), 5 deletions(-)
 
 diff --git a/config/zfs-meta.m4 b/config/zfs-meta.m4
-index b3c1befaa..660d8ccb9 100644
+index 20064a0fb..4d5f545ad 100644
 --- a/config/zfs-meta.m4
 +++ b/config/zfs-meta.m4
 @@ -1,9 +1,10 @@
@@ -67,4 +67,4 @@ index b3c1befaa..660d8ccb9 100644
 +		elif test ! -f ".nogitrelease" && git rev-parse --git-dir > /dev/null 2>&1; then
  			_match="${ZFS_META_NAME}-${ZFS_META_VERSION}"
  			_alias=$(git describe --match=${_match} 2>/dev/null)
- 			_release=$(echo ${_alias}|cut -f3- -d'-'|sed 's/-/_/g')
+ 			_release=$(echo ${_alias}|sed "s/${ZFS_META_NAME}//"|cut -f3- -d'-'|tr - _)
diff --git a/debian/patches/0005-Enable-zed-emails.patch b/debian/patches/0005-Enable-zed-emails.patch
index f605a356..a41693a0 100644
--- a/debian/patches/0005-Enable-zed-emails.patch
+++ b/debian/patches/0005-Enable-zed-emails.patch
@@ -9,23 +9,14 @@ behavior of mdadm.
 
 Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
 ---
- cmd/zed/zed.d/zed.rc | 4 ++--
- 1 file changed, 2 insertions(+), 2 deletions(-)
+ cmd/zed/zed.d/zed.rc | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
 
 diff --git a/cmd/zed/zed.d/zed.rc b/cmd/zed/zed.d/zed.rc
-index 1c278b2ef..41c075c09 100644
+index 9ac77f929..672617f54 100644
 --- a/cmd/zed/zed.d/zed.rc
 +++ b/cmd/zed/zed.d/zed.rc
-@@ -15,7 +15,7 @@
- # Email will only be sent if ZED_EMAIL_ADDR is defined.
- # Disabled by default; uncomment to enable.
- #
--#ZED_EMAIL_ADDR="root"
-+ZED_EMAIL_ADDR="root"
- 
- ##
- # Name or path of executable responsible for sending notifications via email;
-@@ -41,7 +41,7 @@
+@@ -41,7 +41,7 @@ ZED_EMAIL_ADDR="root"
  ##
  # Minimum number of seconds between notifications for a similar event.
  #
diff --git a/debian/patches/0008-Add-systemd-unit-for-importing-specific-pools.patch b/debian/patches/0008-Add-systemd-unit-for-importing-specific-pools.patch
index 98aaf020..4e7c288a 100644
--- a/debian/patches/0008-Add-systemd-unit-for-importing-specific-pools.patch
+++ b/debian/patches/0008-Add-systemd-unit-for-importing-specific-pools.patch
@@ -31,7 +31,7 @@ index e4056a92c..030611419 100644
  enable zfs-mount.service
  enable zfs-share.service
 diff --git a/etc/systemd/system/Makefile.am b/etc/systemd/system/Makefile.am
-index c374a52ac..25d1b99d7 100644
+index 5e65e1db4..8e6baeb68 100644
 --- a/etc/systemd/system/Makefile.am
 +++ b/etc/systemd/system/Makefile.am
 @@ -7,6 +7,7 @@ systemdunit_DATA = \
diff --git a/debian/patches/0012-Fix-zvol_open-lock-inversion.patch b/debian/patches/0012-Fix-zvol_open-lock-inversion.patch
deleted file mode 100644
index eb74550f..00000000
--- a/debian/patches/0012-Fix-zvol_open-lock-inversion.patch
+++ /dev/null
@@ -1,212 +0,0 @@
-From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
-From: Brian Behlendorf <behlendorf1@llnl.gov>
-Date: Fri, 17 Dec 2021 09:52:13 -0800
-Subject: [PATCH] Fix zvol_open() lock inversion
-
-When restructuring the zvol_open() logic for the Linux 5.13 kernel
-a lock inversion was accidentally introduced.  In the updated code
-the spa_namespace_lock is now taken before the zv_suspend_lock
-allowing the following scenario to occur:
-
-    down_read <=== waiting for zv_suspend_lock
-    zvol_open <=== holds spa_namespace_lock
-    __blkdev_get
-    blkdev_get_by_dev
-    blkdev_open
-    ...
-
-     mutex_lock <== waiting for spa_namespace_lock
-     spa_open_common
-     spa_open
-     dsl_pool_hold
-     dmu_objset_hold_flags
-     dmu_objset_hold
-     dsl_prop_get
-     dsl_prop_get_integer
-     zvol_create_minor
-     dmu_recv_end
-     zfs_ioc_recv_impl <=== holds zv_suspend_lock via zvol_suspend()
-     zfs_ioc_recv
-     ...
-
-This commit resolves the issue by moving the acquisition of the
-spa_namespace_lock back to after the zv_suspend_lock which restores
-the original ordering.
-
-Additionally, as part of this change the error exit paths were
-simplified where possible.
-
-Reviewed-by: Tony Hutter <hutter2@llnl.gov>
-Reviewed-by: Rich Ercolani <rincebrain@gmail.com>
-Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
-Closes #12863
-(cherry picked from commit 8a02d01e85556bbe3a1c6947bc11b8ef028d4023)
-Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
----
- module/os/linux/zfs/zvol_os.c | 121 ++++++++++++++++------------------
- 1 file changed, 58 insertions(+), 63 deletions(-)
-
-diff --git a/module/os/linux/zfs/zvol_os.c b/module/os/linux/zfs/zvol_os.c
-index 44caadd58..69479b3f7 100644
---- a/module/os/linux/zfs/zvol_os.c
-+++ b/module/os/linux/zfs/zvol_os.c
-@@ -496,8 +496,7 @@ zvol_open(struct block_device *bdev, fmode_t flag)
- {
- 	zvol_state_t *zv;
- 	int error = 0;
--	boolean_t drop_suspend = B_TRUE;
--	boolean_t drop_namespace = B_FALSE;
-+	boolean_t drop_suspend = B_FALSE;
- #ifndef HAVE_BLKDEV_GET_ERESTARTSYS
- 	hrtime_t timeout = MSEC2NSEC(zvol_open_timeout_ms);
- 	hrtime_t start = gethrtime();
-@@ -517,7 +516,36 @@ retry:
- 		return (SET_ERROR(-ENXIO));
- 	}
- 
--	if (zv->zv_open_count == 0 && !mutex_owned(&spa_namespace_lock)) {
-+	mutex_enter(&zv->zv_state_lock);
-+	/*
-+	 * Make sure zvol is not suspended during first open
-+	 * (hold zv_suspend_lock) and respect proper lock acquisition
-+	 * ordering - zv_suspend_lock before zv_state_lock
-+	 */
-+	if (zv->zv_open_count == 0) {
-+		if (!rw_tryenter(&zv->zv_suspend_lock, RW_READER)) {
-+			mutex_exit(&zv->zv_state_lock);
-+			rw_enter(&zv->zv_suspend_lock, RW_READER);
-+			mutex_enter(&zv->zv_state_lock);
-+			/* check to see if zv_suspend_lock is needed */
-+			if (zv->zv_open_count != 0) {
-+				rw_exit(&zv->zv_suspend_lock);
-+			} else {
-+				drop_suspend = B_TRUE;
-+			}
-+		} else {
-+			drop_suspend = B_TRUE;
-+		}
-+	}
-+	rw_exit(&zvol_state_lock);
-+
-+	ASSERT(MUTEX_HELD(&zv->zv_state_lock));
-+
-+	if (zv->zv_open_count == 0) {
-+		boolean_t drop_namespace = B_FALSE;
-+
-+		ASSERT(RW_READ_HELD(&zv->zv_suspend_lock));
-+
- 		/*
- 		 * In all other call paths the spa_namespace_lock is taken
- 		 * before the bdev->bd_mutex lock.  However, on open(2)
-@@ -542,84 +570,51 @@ retry:
- 		 * the kernel so the only option is to return the error for
- 		 * the caller to handle it.
- 		 */
--		if (!mutex_tryenter(&spa_namespace_lock)) {
--			rw_exit(&zvol_state_lock);
-+		if (!mutex_owned(&spa_namespace_lock)) {
-+			if (!mutex_tryenter(&spa_namespace_lock)) {
-+				mutex_exit(&zv->zv_state_lock);
-+				rw_exit(&zv->zv_suspend_lock);
- 
- #ifdef HAVE_BLKDEV_GET_ERESTARTSYS
--			schedule();
--			return (SET_ERROR(-ERESTARTSYS));
--#else
--			if ((gethrtime() - start) > timeout)
-+				schedule();
- 				return (SET_ERROR(-ERESTARTSYS));
-+#else
-+				if ((gethrtime() - start) > timeout)
-+					return (SET_ERROR(-ERESTARTSYS));
- 
--			schedule_timeout(MSEC_TO_TICK(10));
--			goto retry;
-+				schedule_timeout(MSEC_TO_TICK(10));
-+				goto retry;
- #endif
--		} else {
--			drop_namespace = B_TRUE;
--		}
--	}
--
--	mutex_enter(&zv->zv_state_lock);
--	/*
--	 * make sure zvol is not suspended during first open
--	 * (hold zv_suspend_lock) and respect proper lock acquisition
--	 * ordering - zv_suspend_lock before zv_state_lock
--	 */
--	if (zv->zv_open_count == 0) {
--		if (!rw_tryenter(&zv->zv_suspend_lock, RW_READER)) {
--			mutex_exit(&zv->zv_state_lock);
--			rw_enter(&zv->zv_suspend_lock, RW_READER);
--			mutex_enter(&zv->zv_state_lock);
--			/* check to see if zv_suspend_lock is needed */
--			if (zv->zv_open_count != 0) {
--				rw_exit(&zv->zv_suspend_lock);
--				drop_suspend = B_FALSE;
-+			} else {
-+				drop_namespace = B_TRUE;
- 			}
- 		}
--	} else {
--		drop_suspend = B_FALSE;
--	}
--	rw_exit(&zvol_state_lock);
--
--	ASSERT(MUTEX_HELD(&zv->zv_state_lock));
- 
--	if (zv->zv_open_count == 0) {
--		ASSERT(RW_READ_HELD(&zv->zv_suspend_lock));
- 		error = -zvol_first_open(zv, !(flag & FMODE_WRITE));
--		if (error)
--			goto out_mutex;
--	}
- 
--	if ((flag & FMODE_WRITE) && (zv->zv_flags & ZVOL_RDONLY)) {
--		error = -EROFS;
--		goto out_open_count;
-+		if (drop_namespace)
-+			mutex_exit(&spa_namespace_lock);
- 	}
- 
--	zv->zv_open_count++;
--
--	mutex_exit(&zv->zv_state_lock);
--	if (drop_namespace)
--		mutex_exit(&spa_namespace_lock);
--	if (drop_suspend)
--		rw_exit(&zv->zv_suspend_lock);
--
--	zfs_check_media_change(bdev);
--
--	return (0);
-+	if (error == 0) {
-+		if ((flag & FMODE_WRITE) && (zv->zv_flags & ZVOL_RDONLY)) {
-+			if (zv->zv_open_count == 0)
-+				zvol_last_close(zv);
- 
--out_open_count:
--	if (zv->zv_open_count == 0)
--		zvol_last_close(zv);
-+			error = SET_ERROR(-EROFS);
-+		} else {
-+			zv->zv_open_count++;
-+		}
-+	}
- 
--out_mutex:
- 	mutex_exit(&zv->zv_state_lock);
--	if (drop_namespace)
--		mutex_exit(&spa_namespace_lock);
- 	if (drop_suspend)
- 		rw_exit(&zv->zv_suspend_lock);
- 
--	return (SET_ERROR(error));
-+	if (error == 0)
-+		zfs_check_media_change(bdev);
-+
-+	return (error);
- }
- 
- static void
diff --git a/debian/patches/series b/debian/patches/series
index 8166db91..d2770d39 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -9,4 +9,3 @@
 0009-Patch-move-manpage-arcstat-1-to-arcstat-8.patch
 0010-arcstat-Fix-integer-division-with-python3.patch
 0011-arc-stat-summary-guard-access-to-l2arc-MFU-MRU-stats.patch
-0012-Fix-zvol_open-lock-inversion.patch
diff --git a/upstream b/upstream
index af88d47f..ef83e07d 160000
--- a/upstream
+++ b/upstream
@@ -1 +1 @@
-Subproject commit af88d47f1ee86fae4abcfe0dabca67ea3c94d377
+Subproject commit ef83e07db53e5d1017d3afbf376f4dbb2f6feada
-- 
2.30.2

[pve-devel] [PATCH zfsonlinux 2/2] add zfs-scrub timers+services to zfsutils-linux

[Stoiko Ivanov]

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
---
 debian/zfsutils-linux.install | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/debian/zfsutils-linux.install b/debian/zfsutils-linux.install
index b7501b30..4e4554f8 100644
--- a/debian/zfsutils-linux.install
+++ b/debian/zfsutils-linux.install
@@ -9,6 +9,9 @@ lib/systemd/system/zfs-import@.service
 lib/systemd/system/zfs-import.target
 lib/systemd/system/zfs-import.service
 lib/systemd/system/zfs-mount.service
+lib/systemd/system/zfs-scrub-monthly@.timer
+lib/systemd/system/zfs-scrub-weekly@.timer
+lib/systemd/system/zfs-scrub@.service
 lib/systemd/system/zfs-share.service
 lib/systemd/system/zfs-volume-wait.service
 lib/systemd/system/zfs-volumes.target
-- 
2.30.2

[pve-devel] applied-series: [PATCH zfsonlinux 0/2] update to ZFS 2.1.3

[Thomas Lamprecht]

On 11.03.22 13:55, Stoiko Ivanov wrote:
> updates the submodule to the recently release zfs-2.1.3
> 
> * changeset does not look too exciting (kernel-compatibility fixes, testsuite,
>   FreeBSD)
> * the addition of the systemd timer+service for regular scrubbing is something
>   user-visible (but already present in upstream and our debian packages for a
>   while)
> 
> * the zvol_open lock inversion we cherry-picked is contained in the new tag
> * zed default for mailing has changed upstream:
>   e56dffe4b5cd00df2d46b75a80d3d17c8fe70eec
> * the changes to the other patches were due to some refactorings:
>   745a7f78da95d1969613cd55fcb33df1b46c1e66
> 
> tested with the latest 5.13 kernel on 2 VMs (running ztest)
> 
> Stoiko Ivanov (2):
>   update submodule and patches to ZFS 2.1.3
>   add zfs-scrub timers+services to zfsutils-linux
> 
>  ...META-and-DCH-consistency-in-autoconf.patch |   4 +-
>  debian/patches/0005-Enable-zed-emails.patch   |  17 +-
>  ...md-unit-for-importing-specific-pools.patch |   2 +-
>  .../0012-Fix-zvol_open-lock-inversion.patch   | 212 ------------------
>  debian/patches/series                         |   1 -
>  debian/zfsutils-linux.install                 |   3 +
>  upstream                                      |   2 +-
>  7 files changed, 11 insertions(+), 230 deletions(-)
>  delete mode 100644 debian/patches/0012-Fix-zvol_open-lock-inversion.patch
> 



applied series, thanks!