Hi Heinz
On Thu, Sep 18, 2025 at 9:55 PM Heinz Mauelshagen wrote:
>
> The dm-raid code was using hardcoded integer values to represent the
> read-only/read-write state
> of RAID arrays instead of the proper enumeration constants defined in the
> md_ro_state enumerator type.
>
> Changes:
>
> - R
Hi all
mdadm build fails because of this change.
super0.c: In function ‘update_super0’:
super0.c:672:19: error: ‘mdp_super_t’ {aka ‘struct mdp_superblock_s’}
has no member named ‘recovery_cp’
672 | sb->recovery_cp = 0;
| ^~
Because we recently handled a
Hi Kuai
On Fri, Aug 1, 2025 at 3:11 PM Yu Kuai wrote:
>
> From: Yu Kuai
>
> Redundant data is used to enhance data fault tolerance, and the storage
> method for redundant data vary depending on the RAID levels. And it's
> important to maintain the consistency of redundant data.
>
> Bitmap is use
then IO will still
> have to read all blocks for raid456.
>
> Signed-off-by: Yu Kuai
> Reviewed-by: Christoph Hellwig
> Reviewed-by: Hannes Reinecke
> Reviewed-by: Xiao Ni
> Reviewed-by: Li Nan
> ---
> drivers/md/md-bitmap.h | 1 +
> drivers/md/raid5.c | 6 +
c | 112 +++
> drivers/md/md.h | 4 +-
> drivers/md/raid1-10.c | 2 +-
> drivers/md/raid1.c | 163 +++-
> drivers/md/raid1.h | 22 +-
> drivers/md/raid10.c | 49 ++--
> drivers/md/raid5.c | 30
> 13 files changed, 330 insertions(+), 229 deletions(-)
>
> --
> 2.39.2
>
>
The patch set looks good to me.
Reviewed-by: Xiao Ni
_safe(mddev->kobj.sd, "level");
> diff --git a/drivers/md/md.h b/drivers/md/md.h
> index 67b365621507..d6fba4240f97 100644
> --- a/drivers/md/md.h
> +++ b/drivers/md/md.h
> @@ -796,7 +796,6 @@ struct md_sysfs_entry {
> ssize_t (*show)(struct mddev *, char *);
> ssize_t (*store)(struct mddev *, const char *, size_t);
> };
> -extern const struct attribute_group md_bitmap_group;
>
> static inline struct kernfs_node *sysfs_get_dirent_safe(struct kernfs_node
> *sd, char *name)
> {
> --
> 2.39.2
>
>
Looks good to me.
Reviewed-by: Xiao Ni
}
> }
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index 7ec61ee7b218..999752ec636e 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -8322,7 +8322,7 @@ static int raid5_resize(struct mddev *mddev, sector_t
> sectors)
> mddev->array_sectors > newsize)
> return -EINVAL;
>
> - ret = mddev->bitmap_ops->resize(mddev, sectors, 0, false);
> + ret = mddev->bitmap_ops->resize(mddev, sectors, 0);
> if (ret)
> return ret;
>
> --
> 2.39.2
>
>
Looks good to me.
Reviewed-by: Xiao Ni
在 2025/5/12 上午9:19, Yu Kuai 写道:
From: Yu Kuai
This helper is used to check if blocks can be skipped before calling
into pers->sync_request(), llbiltmap will use this helper to skip
typo error s/llbiltmap/llbitmap/g
resync for unwritten/clean data blocks, and recovery/check/repair for
unw
*/
+ unsigned long normal_io_events; /* IO event timestamp */
atomic_trecovery_active; /* blocks scheduled,
but not written */
wait_queue_head_t recovery_wait;
sector_trecovery_cp;
Looks good to me
Reviewed-by: Xiao Ni
在 2025/4/27 下午5:51, Paul Menzel 写道:
Dear Kuai,
Thank you for your patch.
Am 27.04.25 um 10:29 schrieb Yu Kuai:
From: Yu Kuai
If sync_speed is above speed_min, then is_mddev_idle() will be called
for each sync IO to check if the array is idle, and inflihgt sync_io
infli*gh*t
will be l
On Sun, Apr 27, 2025 at 9:37 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2025/04/22 14:35, Xiao Ni 写道:
> >> + unsigned long last_events;/* IO event
> >> timestamp */
> > Can we use another name? Because mddev has events counter. This name
submit_bio_noacct(bi);
> }
> if (rrdev) {
> - if (s->syncing || s->expanding || s->expanded
> - || s->replacing)
> - md_sync_acct(rrdev->bdev,
> RAID5_STRIPE_SECTORS(conf));
> -
> set_bit(STRIPE_IO_STARTED, &sh->state);
>
> bio_init(rbi, rrdev->bdev, &dev->rvec, 1, op |
> op_flags);
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index e39c45bc0a97..f3a625b00734 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -182,7 +182,6 @@ struct gendisk {
> struct list_head slave_bdevs;
> #endif
> struct timer_rand_state *random;
> - atomic_t sync_io; /* RAID */
> struct disk_events *ev;
>
> #ifdef CONFIG_BLK_DEV_ZONED
> --
> 2.39.2
>
Looks good to me. Reviewed-by: Xiao Ni
On Fri, Apr 18, 2025 at 9:17 AM Yu Kuai wrote:
>
> From: Yu Kuai
>
> If sync_speed is above speed_min, then is_mddev_idle() will be called
> for each sync IO to check if the array is idle, and inflihgt sync_io
> will be limited if the array is not idle.
>
> However, while mkfs.ext4 for a large ra
io_sectors * sync_io_depth(mddev);
> +}
> +
> #define SYNC_MARKS 10
> #defineSYNC_MARK_STEP (3*HZ)
> #define UPDATE_FREQUENCY (5*60*HZ)
> @@ -9195,7 +9265,8 @@ void md_do_sync(struct md_thread *thread)
> msleep(500);
> goto repeat;
> }
> - if (!is_mddev_idle(mddev, 0)) {
> + if (!sync_io_within_limit(mddev) &&
> + !is_mddev_idle(mddev, 0)) {
> /*
> * Give other IO more of a chance.
> * The faster the devices, the less we wait.
> diff --git a/drivers/md/md.h b/drivers/md/md.h
> index 9d55b4630077..b57842188f18 100644
> --- a/drivers/md/md.h
> +++ b/drivers/md/md.h
> @@ -484,6 +484,7 @@ struct mddev {
> /* if zero, use the system-wide default */
> int sync_speed_min;
> int sync_speed_max;
> + int sync_io_depth;
>
> /* resync even though the same disks are shared among md-devices */
> int parallel_resync;
> --
> 2.39.2
>
Looks good to me, reviewed-by: Xiao Ni
* takeover/stop are
> not safe
>*/
> - struct gendisk *gendisk;
> + struct gendisk *gendisk; /* mdraid gendisk */
> + struct gendisk *dm_gendisk; /* dm-raid gendisk */
>
> struct kobject kobj;
> int hold_active;
> --
> 2.39.2
>
Looks good to me, reviewed-by: Xiao Ni
On Fri, Nov 15, 2024 at 4:45 PM Christian Theune wrote:
>
> Hi,
>
> > On 15. Nov 2024, at 09:07, Xiao Ni wrote:
> >
> > On Thu, Nov 14, 2024 at 11:07 PM Christian Theune
> > wrote:
> >>
> >> Hi,
> >>
> >> just a followu
the issue
> >> reliably for about 6 hours now. This is longer than it worked before.
> >> I’m leaving the office for today and will leave things running over night
> >> and report back tomorrow.
> >>
> >> Christian
> >>
> >>> On 11.
On Thu, Nov 7, 2024 at 3:55 PM Yu Kuai wrote:
>
> Hi!
>
> 在 2024/11/06 14:40, Christian Theune 写道:
> > Hi,
> >
> >> On 6. Nov 2024, at 07:35, Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/11/05 18:15, Christian Theune 写道:
> >>> Hi,
> >>> after about 2 hours it stalled again. Here’s the full block
On Mon, May 20, 2024 at 8:38 PM Su Yue wrote:
>
>
> On Thu 09 May 2024 at 09:18, Yu Kuai
> wrote:
>
> > From: Yu Kuai
> >
> > The new helpers will get current sync_action of the array, will
> > be used
> > in later patches to make code cleaner.
> >
> > Signed-off-by: Yu Kuai
> > ---
> > driver
Hi Kuai
I've tested 07reshape5intr with the latest upstream kernel 15 times
without failure. So it's better to have a try with 07reshape5intr with
your patch set.
Regards
Xiao
On Tue, May 21, 2024 at 11:02 AM Oliver Sang wrote:
>
> hi, Yu Kuai,
>
> On Tue, May 21, 2024 at 10:20:54AM +0800, Y
On Tue, May 14, 2024 at 3:39 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/05/14 14:52, Xiao Ni 写道:
> > On Mon, May 13, 2024 at 5:31 PM Yu Kuai wrote:
> >>
> >> From: Yu Kuai
> >>
> >> The new helpers will get current sync_action of the array, w
在 2024/5/9 上午9:18, Yu Kuai 写道:
From: Yu Kuai
Make code cleaner by replace if else if with switch, and it's more
obvious now what is doning for each sync_action. There are no
Hi Kuai
type error s/doning/doing/g
Regards
Xiao
functional changes.
Signed-off-by: Yu Kuai
---
drivers/md/m
On Mon, May 13, 2024 at 5:31 PM Yu Kuai wrote:
>
> From: Yu Kuai
>
> The new helpers will get current sync_action of the array, will be used
> in later patches to make code cleaner.
>
> Signed-off-by: Yu Kuai
> ---
> drivers/md/md.c | 64 +
> driv
On Tue, May 14, 2024 at 2:16 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/05/14 13:51, Xiao Ni 写道:
> > On Mon, May 13, 2024 at 9:57 AM Yu Kuai wrote:
> >>
> >> From: Yu Kuai
> >>
> >> Currently there are lots of flags and the names are confusi
On Mon, May 13, 2024 at 6:19 PM Yu Kuai wrote:
>
> From: Yu Kuai
>
> In order to make code related to sync_thread cleaner in following
> patches, also add detail comment about each sync action.
>
> Signed-off-by: Yu Kuai
> ---
> drivers/md/md.h | 57 +
On Mon, May 13, 2024 at 9:57 AM Yu Kuai wrote:
>
> From: Yu Kuai
>
> Currently there are lots of flags and the names are confusing, since
> there are two main types of flags, sync thread runnng status and sync
> thread action, rearrange and update comment to improve code readability,
> there are
On Mon, Mar 4, 2024 at 4:27 PM Xiao Ni wrote:
>
> On Mon, Mar 4, 2024 at 9:25 AM Xiao Ni wrote:
> >
> > On Mon, Mar 4, 2024 at 9:24 AM Yu Kuai wrote:
> > >
> > > Hi,
> > >
> > > 在 2024/03/04 9:07, Yu Kuai 写道:
> > > > H
On Mon, Mar 4, 2024 at 9:25 AM Xiao Ni wrote:
>
> On Mon, Mar 4, 2024 at 9:24 AM Yu Kuai wrote:
> >
> > Hi,
> >
> > 在 2024/03/04 9:07, Yu Kuai 写道:
> > > Hi,
> > >
> > > 在 2024/03/03 21:16, Xiao Ni 写道:
> > >> Hi all
> &g
On Mon, Mar 4, 2024 at 9:24 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/03/04 9:07, Yu Kuai 写道:
> > Hi,
> >
> > 在 2024/03/03 21:16, Xiao Ni 写道:
> >> Hi all
> >>
> >> There is a error report from lvm regression tests. The case is
> >> lvco
On Sat, Mar 2, 2024 at 6:28 AM Song Liu wrote:
>
> On Fri, Mar 1, 2024 at 7:21 AM Xiao Ni wrote:
> >
> > Hi all
> >
> > This patch set tries to fix dmraid regression problems we face
> > recently. This patch is based on song's md-6.8 branch.
> >
>
k all disks in a stripe_head for reshape
progress")
Signed-off-by: Xiao Ni
---
drivers/md/raid5.c | 22 ++
1 file changed, 10 insertions(+), 12 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 4c1f572cc00f..8d562c1344f4 100644
--- a/drivers/md/raid5.c
d easily by those commands:
while [ 1 ]; do
vgcreate test_vg /dev/loop0 /dev/loop1
lvcreate --type raid1 -L 400M -m 1 -n test_lv test_vg
lvchange -an test_vg
vgremove test_vg -ff
done
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 1
#x27;s keep as small changes as we can. We can rethink about this
in future.
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index db4743ba7f6c..c4624814d94c 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -9496
dmraid regressions.
With patch03 and patch04 and commit 82ec0ae59d02
("md: Make sure md_do_sync() will set MD_RECOVERY_DONE"), all deadlock
problems can be fixed. So revert this one and we can rethink about this
in future.
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 5 +
7;t revert commit 82ec0ae59d02
("md: Make sure md_do_sync() will set MD_RECOVERY_DONE")
It doesn't clear MD_RECOVERY_WAIT before stopping dmraid
Re-write patch01 comment
Xiao Ni (4):
md: Revert "md: Don't register sync_thread for reshape directly"
md: Rev
On Fri, Mar 1, 2024 at 10:38 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/29 23:49, Xiao Ni 写道:
> > This reverts commit ad39c08186f8a0f221337985036ba86731d6aafe.
> >
> > Function stop_sync_thread only wakes up sync task. It also needs to
> > wake up sync threa
On Fri, Mar 1, 2024 at 10:45 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/29 23:49, Xiao Ni 写道:
> > MD_RECOVERY_WAIT is used by dmraid to delay reshape process by patch
> > commit 644e2537fdc7 ("dm raid: fix stripe adding reshape deadlock").
> > Before patch
在 2024/2/26 下午5:36, Yu Kuai 写道:
Hi,
在 2024/02/26 13:12, Xiao Ni 写道:
On Mon, Feb 26, 2024 at 9:31 AM Yu Kuai wrote:
Hi,
在 2024/02/23 21:20, Xiao Ni 写道:
On Fri, Feb 23, 2024 at 11:32 AM Yu Kuai
wrote:
Hi,
在 2024/02/20 23:30, Xiao Ni 写道:
MD_RECOVERY_WAIT is used by dmraid to delay
On Fri, Mar 1, 2024 at 10:12 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/29 23:49, Xiao Ni 写道:
> > Hi all
> >
> > This patch set tries to fix dmraid regression problems when we recently.
> > After talking with Kuai who also sent a patch set which is used to fi
On Fri, Mar 1, 2024 at 7:46 AM Song Liu wrote:
>
> On Thu, Feb 29, 2024 at 2:53 PM Song Liu wrote:
> >
> > On Thu, Feb 29, 2024 at 7:50 AM Xiao Ni wrote:
> > >
> > > This reverts commit 82ec0ae59d02e89164b24c0cc8e4e50de78b5fd6.
> > >
> > >
k all disks in a stripe_head for reshape
progress")
Signed-off-by: Xiao Ni
---
drivers/md/raid5.c | 22 ++
1 file changed, 10 insertions(+), 12 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 8497880135ee..965991a3104f 100644
--- a/drivers/md/raid5.c
d easily by those commands:
while [ 1 ]; do
vgcreate test_vg /dev/loop0 /dev/loop1
lvcreate --type raid1 -L 400M -m 1 -n test_lv test_vg
lvchange -an test_vg
vgremove test_vg -ff
done
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 1
L 16M -n test_lv test_vg
lvconvert -y --stripes 4 /dev/test_vg/test_lv
vgremove test_vg -ff
sleep 1
done
Fixes: 644e2537fdc7 ("dm raid: fix stripe adding reshape deadlock")
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/dm-rai
#x27;s keep as small changes as we can. We can rethink about this
in future.
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 6376b1aad4d9..79dfc015c322 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -9492
This reverts commit 82ec0ae59d02e89164b24c0cc8e4e50de78b5fd6.
The root cause is that MD_RECOVERY_WAIT isn't cleared when stopping raid.
The following patch 'Clear MD_RECOVERY_WAIT when stopping dmraid' fixes
this problem.
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 12 -
This reverts commit ad39c08186f8a0f221337985036ba86731d6aafe.
Function stop_sync_thread only wakes up sync task. It also needs to
wake up sync thread. This problem will be fixed in the following
patch.
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 5 +
drivers/md/raid10.c | 16
shell/dmsetup-integrity-keys.sh
shell/lvresize-fs-crypt.sh
shell/pvck-dump.sh
shell/select-report.sh
And lvconvert-raid-reshape.sh can fail sometimes. But it fails in 6.6
kernel too. So it can return back to the same state with 6.6 kernel.
Xiao Ni (6):
Revert "md: Don't register sync_thr
在 2024/2/1 下午5:25, Yu Kuai 写道:
From: Yu Kuai
md_start_sync() will suspend the array if there are spares that can be
added or removed from conf, however, if reshape is still in progress,
Hi Kuai
Why md_start_sync can run when reshape is still in progress?
md_check_recovery should return w
On Wed, Feb 28, 2024 at 8:44 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/28 20:07, Xiao Ni 写道:
> > I have a question here. Is it the reason sync_thread can't run
> > md_do_sync because kthread_should_stop, so it doesn't have the chance to
> > set MD_RE
在 2024/2/1 下午5:25, Yu Kuai 写道:
From: Yu Kuai
Currently, if reshape is interrupted, then reassemble the array will
register sync_thread directly from pers->run(), in this case
'MD_RECOVERY_RUNNING' is set directly, however, there is no guarantee
that md_do_sync() will be executed, hence stop_s
On Mon, Feb 26, 2024 at 5:36 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/26 13:12, Xiao Ni 写道:
> > On Mon, Feb 26, 2024 at 9:31 AM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/23 21:20, Xiao Ni 写道:
> >>> On Fri, Feb 23, 2024 at
On Mon, Feb 26, 2024 at 9:31 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/23 21:20, Xiao Ni 写道:
> > On Fri, Feb 23, 2024 at 11:32 AM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/20 23:30, Xiao Ni 写道:
> >>> MD_RECOVERY_WAI
On Fri, Feb 23, 2024 at 11:09 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/20 23:30, Xiao Ni 写道:
> > stripe_ahead_of_reshape is used to check if a stripe region cross the
> > reshape position. So first, change the function name to
> > stripe_across_reshape to descr
On Fri, Feb 23, 2024 at 11:06 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/20 23:30, Xiao Ni 写道:
> > If all flush bios finish fast, it doesn't decrease active_io. And it will
> > stuck when stopping array.
> >
> > This can be reproduced by lvm2 test she
On Fri, Feb 23, 2024 at 6:31 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/20 23:30, Xiao Ni 写道:
> > MD_RECOVERY_WAIT is used by dmraid to delay reshape process by patch
> > commit 644e2537fdc7 ("dm raid: fix stripe adding reshape deadlock").
> > Before patch
On Fri, Feb 23, 2024 at 11:32 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/20 23:30, Xiao Ni 写道:
> > MD_RECOVERY_WAIT is used by dmraid to delay reshape process by patch
> > commit 644e2537fdc7 ("dm raid: fix stripe adding reshape deadlock").
> > Before patch
On Wed, Feb 21, 2024 at 1:45 PM Benjamin Marzinski wrote:
>
> On Tue, Feb 20, 2024 at 11:30:55PM +0800, Xiao Ni wrote:
> > Hi all
> >
> > Sorry, I know this patch set conflict with Yu Kuai's patch set. But
> > I have to send out this patch set. Now we
k all disks in a stripe_head for reshape
progress")
Signed-off-by: Xiao Ni
---
drivers/md/raid5.c | 22 ++
1 file changed, 10 insertions(+), 12 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 8497880135ee..4c71df4e2370 100644
--- a/drivers/md/raid5.c
If all flush bios finish fast, it doesn't decrease active_io. And it will
stuck when stopping array.
This can be reproduced by lvm2 test shell/integrity-caching.sh.
But it can't reproduce 100%.
Fixes: fa2bbff7b0b4 ("md: synchronize flush io with array reconfiguration")
Si
d easily by those commands:
while [ 1 ]; do
vgcreate test_vg /dev/loop0 /dev/loop1
lvcreate --type raid1 -L 400M -m 1 -n test_lv test_vg
lvchange -an test_vg
vgremove test_vg -ff
done
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 1
L 16M -n test_lv test_vg
lvconvert -y --stripes 4 /dev/test_vg/test_lv
vgremove test_vg -ff
sleep 1
done
Fixes: 644e2537fdc7 ("dm raid: fix stripe adding reshape deadlock")
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/dm-rai
m. The third one fixes active_io counter bug.
The fouth one fixes the raid5 reshape deadlock problem.
I have run lvm2 regression test. There are 4 failed cases:
shell/dmsetup-integrity-keys.sh
shell/lvresize-fs-crypt.sh
shell/pvck-dump.sh
shell/select-report.sh
Xiao Ni (4):
Clear MD_RECOVERY_WA
On Mon, Feb 19, 2024 at 3:53 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/19 15:27, Xiao Ni 写道:
> > On Sun, Feb 18, 2024 at 2:34 PM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/18 12:53, Xiao Ni 写道:
> >>> Hi Kuai
> >>&g
On Sun, Feb 18, 2024 at 2:34 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 12:53, Xiao Ni 写道:
> > Hi Kuai
> >
> > On Thu, Feb 1, 2024 at 5:30 PM Yu Kuai wrote:
> >>
> >> From: Yu Kuai
> >>
> >> 1) The flag MD_RECOVERY_FROZEN doe
On Sun, Feb 18, 2024 at 4:48 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 16:07, Xiao Ni 写道:
> > On Sun, Feb 18, 2024 at 2:22 PM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/18 13:07, Xiao Ni 写道:
> >>> On Sun, Feb 18, 2024 at
L 16M -n test_lv test_vg
lvconvert -y --stripes 4 /dev/test_vg/test_lv
vgremove test_vg -ff
sleep 1
done
Fixes: 644e2537fdc7 ("dm raid: fix stripe adding reshape deadlock")
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/dm-rai
If all flush bios finish fast, it doesn't decrease active_io. And it will
stuck when stopping array.
This can be reproduced by lvm2 test shell/integrity-caching.sh.
But it can't reproduce 100%.
Fixes: fa2bbff7b0b4 ("md: synchronize flush io with array reconfiguration")
Si
d easily by those commands:
while [ 1 ]; do
vgcreate test_vg /dev/loop0 /dev/loop1
lvcreate --type raid1 -L 400M -m 1 -n test_lv test_vg
lvchange -an test_vg
vgremove test_vg -ff
done
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Signed-off-by: Xiao Ni
---
drivers/md/md.c | 1
lem. The third one fixes active_io counter bug.
I have run lvm2 regression test. lvconvert-raid-reshape.sh is failed. This
patch set doesn't plan to fix it. Kuai's patch set has a patch which should
fix it. And there are other 4 failed cases:
shell/dmsetup-integrity-keys.sh
shell/lvresize-
On Sun, Feb 18, 2024 at 2:51 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 13:56, Xiao Ni 写道:
> > On Thu, Feb 1, 2024 at 5:30 PM Yu Kuai wrote:
> >>
> >> From: Yu Kuai
> >>
> >> stop_sync_thread() will interrupt md_do_sync(), and md_do_sy
On Sun, Feb 18, 2024 at 2:22 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 13:07, Xiao Ni 写道:
> > On Sun, Feb 18, 2024 at 11:24 AM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/18 11:15, Xiao Ni 写道:
> >>> On Sun, Feb 18, 2024 at
On Thu, Feb 1, 2024 at 5:30 PM Yu Kuai wrote:
>
> From: Yu Kuai
>
> stop_sync_thread() will interrupt md_do_sync(), and md_do_sync() must
> set MD_RECOVERY_DONE, so that follow up md_check_recovery() will
> unregister sync_thread, clear MD_RECOVERY_RUNNING and wake up
> stop_sync_thread().
>
> If
On Sun, Feb 18, 2024 at 11:24 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 11:15, Xiao Ni 写道:
> > On Sun, Feb 18, 2024 at 10:34 AM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/18 10:27, Xiao Ni 写道:
> >>> On Sun, Feb 18, 2024 at
Hi Kuai
On Thu, Feb 1, 2024 at 5:30 PM Yu Kuai wrote:
>
> From: Yu Kuai
>
> 1) The flag MD_RECOVERY_FROZEN doesn't mean that sync thread is frozen,
>it only prevent new sync_thread to start, and it can't stop the
>running sync thread;
Agree with this
> 2) The flag MD_RECOVERY_FROZEN do
On Sun, Feb 18, 2024 at 10:34 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 10:27, Xiao Ni 写道:
> > On Sun, Feb 18, 2024 at 9:46 AM Yu Kuai wrote:
> >>
> >> Hi,
> >>
> >> 在 2024/02/18 9:33, Xiao Ni 写道:
> >>> The deadlock problem menti
On Sun, Feb 18, 2024 at 9:46 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/18 9:33, Xiao Ni 写道:
> > The deadlock problem mentioned in this patch should not be right?
>
> No, I think it's right. Looks like you are expecting other problems,
> like mentioned in patch 6, to
On Sun, Feb 18, 2024 at 9:15 AM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/02/16 14:58, Xiao Ni 写道:
> > On Thu, Feb 1, 2024 at 5:30 PM Yu Kuai wrote:
> >>
> >> From: Yu Kuai
> >>
> >> mddev_suspend() never stop sync_thread, hence it doesn't mak
On Thu, Feb 1, 2024 at 5:30 PM Yu Kuai wrote:
>
> From: Yu Kuai
>
> mddev_suspend() never stop sync_thread, hence it doesn't make sense to
> ignore suspended array in md_check_recovery(), which might cause
> sync_thread can't be unregistered.
>
> After commit f52f5c71f3d4 ("md: fix stopping sync
On Wed, Jan 31, 2024 at 9:25 AM Yu Kuai wrote:
>
> Hi, Xiao Ni!
>
> 在 2024/01/31 8:29, Xiao Ni 写道:
> > In my environment, the lvm2 regression test has passed. There are only
> > three failed cases which also fail in kernel 6.6.
> >
> > ### failed: [ndev
On Tue, Jan 30, 2024 at 10:23 AM Yu Kuai wrote:
>
> From: Yu Kuai
>
> Changes in v4:
> - add patch 10 to fix a raid456 deadlock(for both md/raid and dm-raid);
> - add patch 13 to wait for inflight IO completion while removing dm
> device;
>
> Changes in v3:
> - fix a problem in patch 5;
> -
On Fri, Jan 26, 2024 at 8:14 AM Song Liu wrote:
>
> Hi Xiao,
>
> On Thu, Jan 25, 2024 at 5:33 AM Xiao Ni wrote:
> >
> > Hi all
> >
> > I build the kernel 6.7.0-rc8 with this patch set. The lvm2 regression
> > test result:
>
> I believe the patch
Hi all
I build the kernel 6.7.0-rc8 with this patch set. The lvm2 regression
test result:
### failed: [ndev-vanilla] shell/integrity.sh
### failed: [ndev-vanilla] shell/lvchange-partial-raid10.sh
### failed: [ndev-vanilla] shell/lvchange-raid-transient-failures.sh
### faile
On Thu, Jan 25, 2024 at 7:42 PM Yu Kuai wrote:
>
> Hi,
>
> 在 2024/01/25 19:35, Xiao Ni 写道:
> > Hi all
> >
> > This is the result of lvm2 tests:
> > make check
> > ### 426 tests: 319 passed, 74 skipped, 0 timed out, 5 warned, 28
> > failed in 56:0
Hi all
This is the result of lvm2 tests:
make check
### 426 tests: 319 passed, 74 skipped, 0 timed out, 5 warned, 28
failed in 56:04.914
make[1]: *** [Makefile:138: check] Error 1
make[1]: Leaving directory '/root/lvm2/test'
make: *** [Makefile:89: check] Error 2
Do you know where to check whic
On Wed, Jan 24, 2024 at 5:19 PM Yu Kuai wrote:
>
> The new heleprs will be used in dm-raid in later patches to fix
> regressions and prevent calling md_reap_sync_thread() directly.
>
> Signed-off-by: Yu Kuai
> ---
> drivers/md/md.c | 41 +
> drivers/md/md.
On Wed, Jan 24, 2024 at 8:19 PM Xiao Ni wrote:
>
> On Wed, Jan 24, 2024 at 5:18 PM Yu Kuai wrote:
> >
> > First regression related to stop sync thread:
> >
> > The lifetime of sync_thread is designed as following:
> >
> > 1) Decide want to start sync_t
On Wed, Jan 24, 2024 at 5:18 PM Yu Kuai wrote:
>
> First regression related to stop sync thread:
>
> The lifetime of sync_thread is designed as following:
>
> 1) Decide want to start sync_thread, set MD_RECOVERY_NEEDED, and wake up
> daemon thread;
> 2) Daemon thread detect that MD_RECOVERY_NEEDED
Hi all
MD_RECOVERY_WAIT was introduced in d5d885fd5. From this patch,
MD_RECOVERY_WAIT only has one usage during creating raid device.
resync job needs to wait until pers->start finishes(The only place
which is checked). If we remove it from md_do_sync, will it break the
logic? Or we don't need th
On Wed, Nov 29, 2023 at 10:10 AM Eric Wheeler
wrote:
>
> Hi Joe,
>
> I'm not sure who else to CC on this issue, feel free to add others.
>
> Recently we tried putting dm-integrity on NVMe's under MD RAID1 with
> dm-thin metadata (tmeta) on that raid1 mirror (Linux v6.5.7). It worked
> fine for ~1
89 matches
Mail list logo
