From: yangerkun
Once there is a heavy IO load, so many encrypt/decrypt work will occupy
all of the cpu, which may lead the poor performance for other service.
So the idea like 'a2b8b2d97567 ("dm crypt: export sysfs of kcryptd
workqueue")' said seems necessary. We can export
在 2023/7/10 20:22, yangerkun 写道:
Hi!
Thanks a lot for your reminder! There is a weekend and it also take me
some time to understand what happened recently in workqueue[1][2](I have
a limited understanding of workqueue, so the following statement may not
be correct...), sorry for the late
Linus's latest
linux.git. Then we (Mikulas, you, and/or I) can take a closer look at
addressing Tejun's DM_CRYPT_SAME_CPU suggestion.
Fine, I will do it later, and thanks again for your reply!
Thanks,
Yang Erkun.
Thanks,
Mike
On Mon, Jun 26 2023 at 4:43P -0400,
yangerkun wrote:
Hi, Mi
Hi, Mike,
Sorry for the noise. This patch has been proposed for a long time. I
wonder does there any suggestion for the patch. Looking forward to your
reply!
Thanks,
Yang Erkun.
在 2023/3/25 9:01, yangerkun 写道:
Ping...
在 2023/3/1 11:29, yangerkun 写道:
From: yangerkun
Once there is a
Ping...
在 2023/3/1 11:29, yangerkun 写道:
From: yangerkun
Once there is a heavy IO load, so many encrypt/decrypt work will occupy
all of the cpu, which may lead the poor performance for other service.
So the idea like 'a2b8b2d97567 ("dm crypt: export sysfs of kcryptd
workqueue&quo
在 2023/3/7 9:30, yangerkun 写道:
在 2023/3/7 3:01, Mikulas Patocka 写道:
On Mon, 6 Mar 2023, Mikulas Patocka wrote:
Hi,
Thanks a lot for your review!
It's ok to fix the softlockup, but for async write encrypt,
kcryptd_crypt_write_io_submit will add bio to write_tree, and once we
在 2023/3/7 3:01, Mikulas Patocka 写道:
On Mon, 6 Mar 2023, Mikulas Patocka wrote:
Hi,
Thanks a lot for your review!
It's ok to fix the softlockup, but for async write encrypt,
kcryptd_crypt_write_io_submit will add bio to write_tree, and once we
call cond_resched before every kcryptd_io_wri
] dmcrypt_write+0x1e4/0x208 [dm_crypt]
[ 3391.416005][ C12] kthread+0x130/0x138
[ 3391.419911][ C12] ret_from_fork+0x10/0x18
Signed-off-by: Mikulas Patocka
Reported-by: yangerkun
Fixes: dc2676210c42 ("dm crypt: offload writes to thread")
Cc: sta...@vger.kernel.org
---
drivers/md/
From: yangerkun
We meet a softlockup:
localhost login: [ 3391.153255][ C12] watchdog: BUG: soft lockup - CPU#12
stuck for 23s! [dmcrypt_write/2:2897]
...
[ 3391.258372][ C12] CPU: 12 PID: 2897 Comm: dmcrypt_write/2 Tainted: G
L5.10.0+ #8
[ 3391.267288][ C12] Hardware
From: yangerkun
Once there is a heavy IO load, so many encrypt/decrypt work will occupy
all of the cpu, which may lead the poor performance for other service.
So the idea like 'a2b8b2d97567 ("dm crypt: export sysfs of kcryptd
workqueue")' said seems necessary. We can export
在 2023/2/28 2:06, Mike Snitzer 写道:
On Mon, Feb 27 2023 at 1:03P -0500,
Mike Snitzer wrote:
On Mon, Feb 27 2023 at 12:55P -0500,
Mike Snitzer wrote:
On Sun, Feb 26 2023 at 8:31P -0500,
yangerkun wrote:
在 2023/2/26 10:01, Bart Van Assche 写道:
On 2/22/23 19:19, yangerkun wrote
在 2023/2/28 1:55, Mike Snitzer 写道:
On Sun, Feb 26 2023 at 8:31P -0500,
yangerkun wrote:
在 2023/2/26 10:01, Bart Van Assche 写道:
On 2/22/23 19:19, yangerkun wrote:
@@ -1924,6 +1926,10 @@ static int dmcrypt_write(void *data)
BUG_ON(rb_parent(write_tree.rb_node));
+ if
在 2023/2/28 4:56, Mike Snitzer 写道:
On Wed, Feb 22 2023 at 10:18P -0500,
yangerkun wrote:
From: yangerkun
'a2b8b2d97567 ("dm crypt: export sysfs of kcryptd workqueue")' give us
idea to set specific CPU or limit max_active crypt work. However sysfs
will report
在 2023/2/28 9:25, yangerkun 写道:
在 2023/2/28 1:55, Mike Snitzer 写道:
On Sun, Feb 26 2023 at 8:31P -0500,
yangerkun wrote:
在 2023/2/26 10:01, Bart Van Assche 写道:
On 2/22/23 19:19, yangerkun wrote:
@@ -1924,6 +1926,10 @@ static int dmcrypt_write(void *data)
BUG_ON(rb_parent
在 2023/2/26 10:01, Bart Van Assche 写道:
On 2/22/23 19:19, yangerkun wrote:
@@ -1924,6 +1926,10 @@ static int dmcrypt_write(void *data)
BUG_ON(rb_parent(write_tree.rb_node));
+ if (time_is_before_jiffies(start_time + HZ)) {
+ schedule();
+ start_time
From: yangerkun
'a2b8b2d97567 ("dm crypt: export sysfs of kcryptd workqueue")' give us
idea to set specific CPU or limit max_active crypt work. However sysfs
will report a warnning and fail 'cryptsetup refresh test' since the
reload will alloc workqueue with same
From: yangerkun
We meet a softlockup:
localhost login: [ 3391.153255][ C12] watchdog: BUG: soft lockup - CPU#12
stuck for 23s! [dmcrypt_write/2:2897]
...
[ 3391.258372][ C12] CPU: 12 PID: 2897 Comm: dmcrypt_write/2 Tainted: G
L5.10.0+ #8
[ 3391.267288][ C12] Hardware
emporarily
revert this feature to fix the problem.
What we actually should do is give a unique name once we try reload
table, we can use ida now.
Signed-off-by: yangerkun
---
drivers/md/dm-crypt.c | 28 +++-
1 file changed, 23 insertions(+), 5 deletions(-)
diff --git a/d
Hi Mike, does there any advice for this patchset?
Thanks,
Kun.
在 2020/6/20 5:10, Bryan Gurney 写道:
From: yangerkun
From: yangerkun
Some type of message(queryblock/countbadblocks/removebadblock) may better
report results to user directly. Do it with DMEMIT.
[Bryan: maintain __func__ output
Hi,
Thanks a lot for your fix and this patchset! It looks better than
before!
Thanks,
Kun.
在 2020/6/20 5:10, Bryan Gurney 写道:
From: yangerkun
From: yangerkun
Some type of message(queryblock/countbadblocks/removebadblock) may better
report results to user directly. Do it with DMEMIT
This interface may help anyone who want to know all badblocks without
query block for each.
Signed-off-by: yangerkun
---
drivers/md/dm-dust.c | 27 +++
1 file changed, 27 insertions(+)
diff --git a/drivers/md/dm-dust.c b/drivers/md/dm-dust.c
index a0c75c104de0
Some type of message(queryblock/countbadblocks/removebadblock) may better
report results to user directly. Do it with DMEMIT.
Signed-off-by: yangerkun
---
drivers/md/dm-dust.c | 31 ++-
1 file changed, 18 insertions(+), 13 deletions(-)
diff --git a/drivers/md/dm
Since we support the listbadblocks command, introduce the detail in the
doc.
Signed-off-by: yangerkun
---
.../admin-guide/device-mapper/dm-dust.rst| 16
1 file changed, 16 insertions(+)
diff --git a/Documentation/admin-guide/device-mapper/dm-dust.rst
b/Documentation
List all bad blocks may help user check the status of dust. We add this
logical in dust_status, which will return result to user directly.
Signed-off-by: yangerkun
---
drivers/md/dm-dust.c | 25 +
1 file changed, 25 insertions(+)
diff --git a/drivers/md/dm-dust.c b
在 2020/6/18 23:40, Mike Snitzer 写道:
On Thu, Jun 18 2020 at 5:06am -0400,
yangerkun wrote:
This patchset will list badblocks while we do 'dmsetup status'.
Image that we have mark block 1 and 2 as bad block, run following
command will list all bad blocks:
$ sudo dmsetup st
Since some type of message will report the results to user directly,
we should update the doc too.
Signed-off-by: yangerkun
---
.../admin-guide/device-mapper/dm-dust.rst | 15 ---
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/Documentation/admin-guide/device
is logical in dust_status instead dust_message since the
result can return to user directly.
yangerkun (2):
dm dust: list badblock in dust_status
dm dust: introduce list badblocks in the rst
.../admin-guide/device-mapper/dm-dust.rst | 14 +++
drivers/md/dm-dust.c
or we will list them multi
line which each line means one bad block.
v2->v3:
Realize this logical in .message, change logical for some type message to
report results to user too.
yangerkun (4):
dm dust: report some message results back to user directly
dm dust: update doc after message
Since we can list bad blocks with command status, introduce the detail
in the doc.
Signed-off-by: yangerkun
---
.../admin-guide/device-mapper/dm-dust.rst | 14 ++
1 file changed, 14 insertions(+)
diff --git a/Documentation/admin-guide/device-mapper/dm-dust.rst
b
在 2020/6/16 20:28, Bryn M. Reeves 写道:
On Mon, Jun 15, 2020 at 06:03:31PM +0100, Alasdair G Kergon wrote:
On Mon, Jun 08, 2020 at 04:45:11PM +0800, yangerkun wrote:
$ sudo dmsetup message dust1 0 listbadblocks
The following message will appear, listing one bad block number per
Did you
Hi,
Sorry for not send to Mike. Can you consider to apply this patchset?
Thanks,
Kun.
在 2020/6/8 16:45, yangerkun 写道:
We can add/remove/query the badblocks, but no interface to list all
badblocks when we want to know the badblocks we ever set still
available. Add message listbadblocks to do
device-mapper: dust: dust_list_badblocks: badblocks list end.
yangerkun (2):
dm dust: add interface to list all badblocks
dm dust: introduce listbadblocks in the rst
.../admin-guide/device-mapper/dm-dust.rst | 18
drivers/md/dm-dust.c | 21
Since we support the listbadblocks command, introduce the detail in the
doc.
Signed-off-by: yangerkun
Reviewed-by: Bryan Gurney
---
.../admin-guide/device-mapper/dm-dust.rst | 18 ++
1 file changed, 18 insertions(+)
diff --git a/Documentation/admin-guide/device-mapper/dm
This interface may help anyone want to know all badblocks without query
block one by one.
Signed-off-by: yangerkun
Reviewed-by: Bryan Gurney
---
drivers/md/dm-dust.c | 21 +
1 file changed, 21 insertions(+)
diff --git a/drivers/md/dm-dust.c b/drivers/md/dm-dust.c
index
在 2020/6/6 3:47, Bryan Gurney 写道:
On Fri, Jun 5, 2020 at 3:48 AM yangerkun wrote:
Since we support the listbadblocks command, introduce the detail in the
doc.
Signed-off-by: yangerkun
Thank you for the documentation. I have a few corrections listed below:
---
.../admin-guide/device
Since we support the listbadblocks command, introduce the detail in the
doc.
Signed-off-by: yangerkun
---
.../admin-guide/device-mapper/dm-dust.rst| 16
1 file changed, 16 insertions(+)
diff --git a/Documentation/admin-guide/device-mapper/dm-dust.rst
b/Documentation
This interface may help anyone want to know all badblocks without query
block one by one.
Signed-off-by: yangerkun
---
drivers/md/dm-dust.c | 21 +
1 file changed, 21 insertions(+)
diff --git a/drivers/md/dm-dust.c b/drivers/md/dm-dust.c
index ff03b90072c5..903c0d158c6e
: badblocks list end.
yangerkun (2):
dm dust: add interface to list all badblocks
dm dust: introduce listbadblocks in the rst
.../admin-guide/device-mapper/dm-dust.rst | 16 ++
drivers/md/dm-dust.c | 21 +++
2 files changed, 37
Hi,
Mike Snitzer wrote on 2019/2/15 2:55:
On Thu, Feb 14 2019 at 4:19am -0500,
yangerkun wrote:
Since 57c36519e4("dm: fix clone_bio() to trigger blk_recount_segments()")
has been reverted by fa8db494("dm: don't use bio_trim() afterall"), the
problem tha
: fa8db4948f522("dm: don't use bio_trim() afterall")
Signed-off-by: yangerkun
---
drivers/md/dm.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 515e6af..b22ac04 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -1336,6 +1336,7 @@
Hi,
Nowdays, i am trying dm-zoned+f2fs to use SMR block device. But, there
is a problem that i can not distinguish it is problem of block device or
the software.
The steps and phenomena of recurrence are as follows:
1.dmzadm --format /dev/sdf
2.echo "0 `blockdev --getsize /dev/sdf` zoned /dev/
41 matches
Mail list logo
