NFS emulates flocks via posix lock on server and fl->fl_owner is set to filp.
Signed-off-by: Stanislav Kinsburskiy
---
fs/locks.c |1 +
1 file changed, 1 insertion(+)
diff --git a/fs/locks.c b/fs/locks.c
index cb7da61..a5ab0c0 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -2503,6 +250
Unlinked files are supported starting from version spfs-0.07.0-1.vz7
27.05.2016 12:09, Stanislav Kinsburskiy пишет:
1. Feature
NFS mount point migration via CRIU
https://jira.sw.ru/browse/PSBM-26967
2. Description
CRIU now supports NFS file system migration (all versions).
CRIU itself does
UNIX sockets are supported starting from version spfs-0.08.0-1.vz7
27.05.2016 12:09, Stanislav Kinsburskiy пишет:
1. Feature
NFS mount point migration via CRIU
https://jira.sw.ru/browse/PSBM-26967
2. Description
CRIU now supports NFS file system migration (all versions).
CRIU itself does
These scripts are obsolete, and their logic is integrated to the source code
starting from version 2.4.0.3.
This means, that libvzctl wtih this patch applied has to depend on criu
2.4.0.3
Signed-off-by: Stanislav Kinsburskiy
---
scripts/vz-rst.in |8
1 file changed, 8 deletions
ec385c6ab4d58f33b100192a96a.
3) This patch is not aimed to fix the issue, but to show the problem root.
Look like this problem moght be applicable to other hunks from the commit,
mentioned above.
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/sched.c |1 -
1 file changed, 1 deletion(-)
d
03.08.2016 19:36, Jeff Layton пишет:
On Wed, 2016-08-03 at 20:54 +0400, Stanislav Kinsburskiy wrote:
Otherwise freezer cgroup state might never become "FROZEN".
Here is a deadlock scheme for 2 processes in one freezer cgroup,
which is
freezing:
CPU 0
04.08.2016 15:16, Jeff Layton пишет:
On Thu, 2016-08-04 at 12:55 +0200, Stanislav Kinsburskiy wrote:
03.08.2016 19:36, Jeff Layton пишет:
On Wed, 2016-08-03 at 20:54 +0400, Stanislav Kinsburskiy wrote:
Otherwise freezer cgroup state might never become "FROZEN".
Here is a deadl
est_wait_answer is called, will lead to request
interruption, producing SIGBUS error in page fault handler (filemap_fault).
https://jira.sw.ru/browse/PSBM-53581
Signed-off-by: Stanislav Kinsburskiy
---
fs/fuse/dev.c | 42 --
1 file changed, 16 in
This patch fixes wrong SIGBUS result in page fault handler for fuse file, when
process received a signal.
https://jira.sw.ru/browse/PSBM-53581
---
Stanislav Kinsburskiy (2):
new helper: wait_event_killable_exclusive()
fuse: handle only fatal signals while waiting request answer
Backport of 6a0fb306738994d6f091791aeb11a5dc87ad8f4c.
Signed-off-by: Stanislav Kinsburskiy
---
include/linux/wait.h | 26 ++
1 file changed, 26 insertions(+)
diff --git a/include/linux/wait.h b/include/linux/wait.h
index 65da9e3..8475f2d 100644
--- a/include/linux
signals to block is present in process pending mask,
and if so - set (!) TIF_SIGPENDING on the task.
IOW, any pending signal remains pending after call to blocks_sigs().
And that's is the root of the issue (as it described in the patch comment).
Thanks,
Maxim
On 10/13/2016 03:03 AM, Stanis
16.10.2016 05:21, Maxim Patlasov пишет:
Stas,
On 10/14/2016 03:30 AM, Stanislav Kinsburskiy wrote:
14.10.2016 02:23, Maxim Patlasov пишет:
Stas,
The series look fine, so:
Acked-by: Maxim Patlasov
But, please, refine the description of the second patch. It must
explain clearly why
Kostya, please, modify patch comment with the following substitution:
s/
it doesn't drop pending signal flag/
it doesn't drop pending signal flag, if processes is traced/
g
Thanks.
13.10.2016 12:03, Stanislav Kinsburskiy пишет:
This patch is b
>
> >IOW, any signal, arrived to the process, which does page fault
> handling on fuse
> >file, _before_ request_wait_answer() is called, will lead to request
> >interruption, producing SIGBUS error in page fault handler
> (filemap_fault).
>
>
> Thanks,
1. Feature
NFS mount point migration via CRIU
https://jira.sw.ru/browse/PSBM-26967
2. Description
CRIU now supports NFS file system migration (all versions).
CRIU itself does mount of SPFS file system (fuse based) instead of NFS
and starts SPFS manager program.
SPFS-manager does NFS remount an
Otherwise migration can fail
https://jira.sw.ru/browse/PSBM-58178
Signed-off-by: Stanislav Kinsburskiy
---
kernel/kmod.c |4
1 file changed, 4 insertions(+)
diff --git a/kernel/kmod.c b/kernel/kmod.c
index 63748d4..5253cae 100644
--- a/kernel/kmod.c
+++ b/kernel/kmod.c
@@ -422,6
28.12.2016 14:49, Dmitry Safonov пишет:
On 12/28/2016 04:37 PM, Stanislav Kinsburskiy wrote:
Otherwise migration can fail
https://jira.sw.ru/browse/PSBM-58178
Signed-off-by: Stanislav Kinsburskiy
Just for check - these modules should be loaded on sys_mount()?
In this case it should be
it's dead
alredy.
Don't see any better solution so far (except investigation, why -ENETUNREACH
is returned).
So let's simply break the loop when it's called from ve_drop_context().
https://jira.sw.ru/browse/PSBM-60905
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/xprt.c
22.02.2017 18:04, Dmitry Safonov пишет:
On 02/22/2017 08:05 PM, Stanislav Kinsburskiy wrote:
Have no idea, why kernel_connect() in xs_tcp_setup_socket() return
-ENETUNREACH
with this call stack:
[root@uqvm098 ~]# cat /proc/14753/stack
[] rpc_wait_bit_killable+0x11/0x60 [sunrpc
This operations can be asynchonious. So, current VE is not the right one.
VE from transport network has to be taken instead.
Fixes commit 4ad21d242066afcb4f55000c308322c3bad98a2b
https://jira.sw.ru/browse/PSBM-62616
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/xprt.c |4 +++-
1
And current VE name and VE net_ns pointer.
https://jira.sw.ru/browse/PSBM-63769
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/xprt.c |9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 1dd1598..c28b3e3 100644
--- a
rowse/PSBM-64015
Notes: this patch obsoletes the following commits:
455d5ba9506998f589b9929266c12627c9baf589
3e48f57a452b597b23039c3278e67788ac48f8d4
ed93afb07e58beb447556adbb25af11b9da1b62c
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/venetdev.c | 32
1
This shouldn't happen. VEIP object is identified by veid, which is unique per
VE structure.
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/veip_mgmt.c |2 ++
drivers/net/venetdev.c |2 +-
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/veip_mgm
Major changes to current behaviour:
1) "SHUTDOWN" hook chain added
2) VEIP object is destroyed in "SHUTDOWN" hook
3) Now VEIP object live cycle is tightly bound to VE object. IOW, it won't be
reused, but EEXIST will be returned instead.
https://jira.sw.ru/browse/PS
It will be used to release VEIP object.
Signed-off-by: Stanislav Kinsburskiy
---
include/linux/ve_proto.h |1 +
kernel/ve/ve.c |2 ++
2 files changed, 3 insertions(+)
diff --git a/include/linux/ve_proto.h b/include/linux/ve_proto.h
index d2dc12d..02f35e2 100644
--- a/include
This one is used only in one place.
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/venetdev.c |3 +--
include/linux/venet.h |1 -
2 files changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/net/venetdev.c b/drivers/net/venetdev.c
index 4a7876f..a4388fd 100644
--- a
al net put).
https://jira.sw.ru/browse/PSBM-64869
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/venetdev.c | 13 -
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/drivers/net/venetdev.c b/drivers/net/venetdev.c
index b7d7596..5e832d3 100644
--- a/drivers/net
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/venetdev.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/drivers/net/venetdev.c b/drivers/net/venetdev.c
index 192ca3c..5a09575 100644
--- a/drivers/net/venetdev.c
+++ b/drivers/net/venetdev.c
@@ -252,13
It takes spinlock only once for both __veip_stop and __venet_ext_clean
(which is converted to unprotected, because now called only from this helper).
New helper is now used in venet_dellink.
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/venetdev.c | 22 ++
1 file
Memmory pressure is referenced as a pointer in socket structure.
Simple check for pointer address will always return true for each socket,
allocated with enabled mem_cg.
Memory pressure value has to checked instead.
https://jira.sw.ru/browse/PSBM-66468
Signed-off-by: Stanislav Kinsburskiy
That's the way to support multiple distributions.
https://jira.sw.ru/browse/PSBM-66901
Signed-off-by: Stanislav Kinsburskiy
---
scripts/nfs-ports-allow.sh | 21 -
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git a/scripts/nfs-ports-allow.sh b/scripts/nfs-
That's the way to support multiple distributions, including those without
$PATH variable being set.
https://jira.sw.ru/browse/PSBM-66967
Signed-off-by: Stanislav Kinsburskiy
---
scripts/systemd-autofs-restart.sh | 28 ++--
1 file changed, 18 insertions(+
https://jira.sw.ru/browse/PSBM-43609
Signed-off-by: Stanislav Kinsburskiy
---
net/netfilter/xt_owner.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/netfilter/xt_owner.c b/net/netfilter/xt_owner.c
index 79558fe..942cce1 100644
--- a/net/netfilter/xt_owner.c
+++ b
It's good enough for us. It won't work properly in case of setting rules by
joining container network namespace without VE cgroup, but it's acceptable,
because proper fix needs a lot of backporting.
https://jira.sw.ru/browse/PSBM-43609
Signed-off-by: Stanislav Kinsburskiy
---
06.06.2017 12:27, Cyrill Gorcunov пишет:
On Tue, Jun 06, 2017 at 01:23:55PM +0300, Cyrill Gorcunov wrote:
On Tue, Jun 06, 2017 at 02:00:32PM +0400, Stanislav Kinsburskiy wrote:
It's good enough for us. It won't work properly in case of setting rules by
joining container network
Signed-off-by: Stanislav Kinsburskiy
---
criu/files-reg.c |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/criu/files-reg.c b/criu/files-reg.c
index bcfe1fd..f749445 100644
--- a/criu/files-reg.c
+++ b/criu/files-reg.c
@@ -219,7 +219,7 @@ again
08.06.2017 16:33, Dmitry Safonov пишет:
On 06/08/2017 05:00 PM, Stanislav Kinsburskiy wrote:
Signed-off-by: Stanislav Kinsburskiy
---
criu/files-reg.c |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/criu/files-reg.c b/criu/files-reg.c
index bcfe1fd..f749445
Regardless its current VE.
https://jira.sw.ru/browse/PSBM-67185
Signed-off-by: Stanislav Kinsburskiy
---
fs/fuse/inode.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 090a7ab..a8d7924 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse
Previously this path had slash at the end. Now it doesn't have.
Update usage then.
Signed-off-by: Stanislav Kinsburskiy
---
criu/files-reg.c |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/criu/files-reg.c b/criu/files-reg.c
index e812b27..2577a48 100644
--- a
SPFS doesn't work with CRIU v3.
This patch set fixes the issue on the CRIU side.
The following series implements...
---
Stanislav Kinsburskiy (2):
spfs: pass both "mountpoint" and "ns_mountpoint" paths to manager
spfs: fix usage of mi->ns_mountpoin
mount
to.
One more difference: "root" is not required anymore. Originally, path to mount
SPFS was constructed as follows:
SPFS_mnt_path = CT_root + mi->ns_mountpoint
Now mounts are created elsewhere and full path is defined in mi->mountpoint.
Signed-off-by: Stanislav Kinsburskiy
https://jira.sw.ru/browse/PSBM-66943
09.06.2017 15:44, Stanislav Kinsburskiy пишет:
SPFS doesn't work with CRIU v3.
This patch set fixes the issue on the CRIU side.
The following series implements...
---
Stanislav Kinsburskiy (2):
spfs: pass both "mountpoint" and
13.06.2017 11:53, Kirill Tkhai пишет:
On 09.06.2017 16:27, Stanislav Kinsburskiy wrote:
Kirill, please, review.
Перенаправленное сообщение
Тема: [Devel] [PATCH] fuse: allow to mount if process is in init_user_ns
Дата: Fri, 9 Jun 2017 17:09:49 +0400
От: Stanislav
This series fixes two issues:
1) Moves "NFS is overmounted" check to the right place
2) Apply "NFS is overmounted" check to all versions
Also, it prints additional information about nested mountpoint
The following series implements...
---
Stanislav Kinsburskiy (2
: Stanislav Kinsburskiy
---
criu/mount.c | 35 ++-
1 file changed, 22 insertions(+), 13 deletions(-)
diff --git a/criu/mount.c b/criu/mount.c
index cc54e68..46ad66c 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -397,6 +397,25 @@ static bool mounts_equal(struct
https://jira.sw.ru/browse/PSBM-67466
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c |5 +
1 file changed, 5 insertions(+)
diff --git a/criu/mount.c b/criu/mount.c
index 46ad66c..0b612a3 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -411,6 +411,11 @@ static bool
: Stanislav Kinsburskiy
---
criu/mount.c | 27 ++-
1 file changed, 22 insertions(+), 5 deletions(-)
diff --git a/criu/mount.c b/criu/mount.c
index 942515e..46ad66c 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -397,6 +397,25 @@ static bool mounts_equal(struct mount_info
Thanks, Kirill!
I'm sorry, but this series is wrong.
I'll send another one soon.
20.06.2017 17:46, Kirill Tkhai пишет:
One minor nit below
On Tue, Jun 20, 2017 at 18:28, Stanislav Kinsburskiy wrote:
Call it on mount points tree building and abort in case NFS is overmounted.
Please, ignore this series
20.06.2017 17:28, Stanislav Kinsburskiy пишет:
This series fixes two issues:
1) Moves "NFS is overmounted" check to the right place
2) Apply "NFS is overmounted" check to all versions
Also, it prints additional information about nested mount
https://jira.sw.ru/browse/PSBM-67466
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c |7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/criu/mount.c b/criu/mount.c
index 667b279..edbd140 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -751,8 +751,13 @@ static
This series fixes "NFS is overmounted" check to be appied for all NFS versions.
Also, it prints additional information about nested mountpoint
The following series implements...
---
Stanislav Kinsburskiy (2):
mounts: add unsupported mount helper
mounts: print NFS nested
Call it on mount points validation and abort in case NFS is overmounted.
This patch also adds nfs_mount helper, which fixes the issue with skipping
this check for NFSv4 mount point.
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c | 23 ---
1 file changed, 20
Intended to be used to abort all SUNRPC operations. For example in case of
fast stop with unreachable network.
Signed-off-by: Stanislav Kinsburskiy
---
include/linux/ve.h |3 +++
kernel/ve/ve.c | 15 ++-
2 files changed, 17 insertions(+), 1 deletion(-)
diff --git a
his solution allows us to mimic hardware behaviour closely. IOW, on normal
stop (or migration) SUNRPC operates normally. But when user wants to stop it
fast (i.e. he doesn't care about data, etc), then he can use "--fast" option
and this patch comes into play.
https://jira.sw.ru/
unreachable
https://jira.sw.ru/browse/PSBM-66510
---
Stanislav Kinsburskiy (2):
ve: add boolean "sunrpc_abort" attribute to VE cgroup
sunrpc: bring back SUNRPC task abort logic
include/linux/ve.h |3 +++
kernel/ve/ve.c | 15 ++-
net/sunrpc/sche
This is needed to break infinite loop in SUNRPC state machine, leading to
unstoppable container in case of unreachable network.
https://jira.sw.ru/browse/PSBM-67544
Note: needs kernel rh7-3.10.0-514.16.1.vz7.32.12 or newer
Signed-off-by: Stanislav Kinsburskiy
---
lib/env_nsops.c | 16
Systemctl may fail. Let's check its result and mountpoint string after call.
https://jira.sw.ru/browse/PSBM-67544
Signed-off-by: Stanislav Kinsburskiy
---
scripts/systemd-autofs-restart.sh | 10 ++
1 file changed, 10 insertions(+)
diff --git a/scripts/systemd-autofs-restart.
Signed-off-by: Stanislav Kinsburskiy
---
scripts/systemd-autofs-restart.sh | 11 ---
1 file changed, 4 insertions(+), 7 deletions(-)
diff --git a/scripts/systemd-autofs-restart.sh
b/scripts/systemd-autofs-restart.sh
index d39e0cd..72d2ef7 100755
--- a/scripts/systemd-autofs
carefully
https://jira.sw.ru/browse/PSBM-67544
The following series implements...
---
Stanislav Kinsburskiy (2):
scripts: check service mountpoint is not empty in
systemd-autofs-restart.sh
scripts: cleanup error output in systemd-autofs-restart.sh
scripts/systemd-autofs
Indeed. Sorry about this. My bad.
23.06.2017 22:10, Dmitry Safonov пишет:
On 06/23/2017 06:38 PM, Stanislav Kinsburskiy wrote:
carefully
https://jira.sw.ru/browse/PSBM-67544
The following series implements...
---
Stanislav Kinsburskiy (2):
scripts: check service mountpoint is not
Please, drop this patch.
I'll sent next version.
23.06.2017 15:42, Stanislav Kinsburskiy пишет:
This is needed to break infinite loop in SUNRPC state machine, leading to
unstoppable container in case of unreachable network.
https://jira.sw.ru/browse/PSBM-67544
Note: needs kernel rh7-3
quot;.
This solution allows us to mimic hardware behaviour closely. IOW, on normal
stop (or migration) SUNRPC operates normally. But when user wants to stop it
fast (i.e. he doesn't care about data, etc), then he can use "--fast" option
and this patch comes into play.
https://jira.sw.ru/
This patch aborts SUNRPC tasks upon raise of per-net "kill_tasks" toggle.
This is needed to abort pernding tasks, which are waiting for timeout.
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/clnt.c | 18 ++
1 file changed, 18 insertions(+)
diff --git a/net/sun
Boolean per-net SUNPRC proc entry. Will be used to abort SUNRPC traffic when
necessary.
Signed-off-by: Stanislav Kinsburskiy
---
include/linux/sunrpc/clnt.h |4 ++
net/sunrpc/clnt.c | 71 +++
net/sunrpc/netns.h |3 ++
net
https://jira.sw.ru/browse/PSBM-66510
---
Stanislav Kinsburskiy (3):
sunrpc: add "kill-tasks" proc entry
sunrpc: bring back SUNRPC task abort logic
sunrpc: pending tasks kill helper added
include/linux/sunrpc/clnt.h |6 +++
net/sunrpc/clnt.c
it process is used
rather than VE cgroup handle.
Signed-off-by: Stanislav Kinsburskiy
---
lib/env_nsops.c | 52
1 file changed, 48 insertions(+), 4 deletions(-)
diff --git a/lib/env_nsops.c b/lib/env_nsops.c
index d62dff4..74d6402 100644
NFS is replaced by FUSE-based SPFS on restore. And SPFS is mounted via
libfuse, which uses "MS_NODEV" mount option.
Thus it's impossible to open a device file on SPFS.
This patch forbids dumping to prevent restore failure.
https://jira.sw.ru/browse/PSBM-55590
Signed-off-by: Stanis
All the mounts on restore stage have to be supported.
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/criu/mount.c b/criu/mount.c
index edbd140..6916974 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -834,7 +834,7
Sure!
Thanks a lot for the hint!
28 июня 2017 г. 6:47 PM пользователь Kirill Tkhai
написал:
On Wed, Jun 28, 2017 at 15:01, Stanislav Kinsburskiy wrote:
> All the mounts on restore stage have to be supported.
>
> Signed-off-by: Stanislav Kinsburskiy
> ---
> criu/mount.c |
er so far and migration
attempt can lead to different issues after restore is completed by CRIU.
https://jira.sw.ru/browse/PSBM-66945
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c | 67 +-
1 file changed, 66 insertions(+), 1 del
ve bottleneck when many processes are concurrently trying to
lock large file like SQL database.
Ported from rhel6.
https://jira.sw.ru/browse/PSBM-66572
Signed-off-by: Stanislav Kinsburskiy
---
fs/nfsd/nfs3proc.c |3 +++
fs/nfsd/nfs3xdr.c |9 +++--
fs/nfsd/xdr3.h |1 +
3 f
The following series implements...
---
Stanislav Kinsburskiy (2):
netlink: fix warning print
sk-queue: fix non-fatal print log level
criu/sk-netlink.c |2 +-
criu/sk-queue.c |6 +++---
2 files changed, 4 insertions(+), 4 deletions
Signed-off-by: Stanislav Kinsburskiy
---
criu/sk-queue.c |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/criu/sk-queue.c b/criu/sk-queue.c
index 8ab53f6..5e2601c 100644
--- a/criu/sk-queue.c
+++ b/criu/sk-queue.c
@@ -230,9 +230,9 @@ static int dump_sk_creds(struct
Signed-off-by: Stanislav Kinsburskiy
---
criu/sk-netlink.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/criu/sk-netlink.c b/criu/sk-netlink.c
index 9933e10..a1b09e6 100644
--- a/criu/sk-netlink.c
+++ b/criu/sk-netlink.c
@@ -134,7 +134,7 @@ static int dump_nl_opts(int sk
Igor, could you please review?
26.06.2017 17:04, Stanislav Kinsburskiy пишет:
This is needed to break infinite loop in SUNRPC state machine, leading to
unstoppable container in case of unreachable network.
https://jira.sw.ru/browse/PSBM-67544
Note: needs kernel rh7-3.10.0-514.16.1.vz7.32.12
igned-off-by: Stanislav Kinsburskiy
---
kernel/sys.c |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/kernel/sys.c b/kernel/sys.c
index 9a681ae..f8f1dd9 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -2329,12 +2329,12 @@ static int prctl_set_mm(int opt, unsigned long
This is needed to make iotop work in container:
https://jira.sw.ru/browse/PSBM-56171
Signed-off-by: Stanislav Kinsburskiy
---
kernel/taskstats.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 37d4456..ae620de 100644
--- a
And print kill error only if it's set.
This is a precursor patch, needed to support tasks killing in thawed
container.
Signed-off-by: Stanislav Kinsburskiy
---
lib/env_nsops.c |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/env_nsops.c b/lib/env_nsops.c
unreachable network
The solution is to kill spfs process before freezing attempt on fast stop
stage. And since the spfs pid in unknown, this is implemented by killing
container processes before freezing (error are not handled).
https://jira.sw.ru/browse/PSBM-68232
---
Stanislav Kinsburskiy (2
l the container tasks without freezing (and optimization), or at least
release processes, waiting for spfs reply in kernel.
https://jira.sw.ru/browse/PSBM-68232
Signed-off-by: Stanislav Kinsburskiy
---
lib/env_nsops.c |4
1 file changed, 4 insertions(+)
diff --git a/lib/env_nsops.c b/lib/en
There could be a situation, when RPC task is not yet sleeping, but already
passed the check for "kill_tasks" in __rpc_execute().
There is a race in such a case, because task will fall into sleep anyway.
This patch makes this probability a little bit less.
Signed-off-by: Stanislav K
rowse/PSBM-68214
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/sched.c |6 ++
1 file changed, 6 insertions(+)
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index cd41946..a55 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -354,6 +354,12 @@ static void __
This helper is much better.
First, is's in upsteam (I wrote it, but forgot :( ).
Second, it kills *all* tasks (including sleeping in other than "pending"
queues).
Signed-off-by: Stanislav Kinsburskiy
---
net/sunrpc/clnt.c |7 ++-
1 file changed, 2 insertions(+), 5 del
This series fix two races with SUNRPC tasks killing on fast stop and also fix
the kill itself (*all* the tasks are killed now, but not only those in
"pending" wait queue).
https://jira.sw.ru/browse/PSBM-68214
---
Stanislav Kinsburskiy (3):
sunrpc: fix race between "kill
ectly.
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c | 79 +-
1 file changed, 78 insertions(+), 1 deletion(-)
diff --git a/criu/mount.c b/criu/mount.c
index 6916974..26f3aa5 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -746,6 +7
This is needed to make lsmod happy.
https://jira.sw.ru/browse/PSBM-63892
---
Stanislav Kinsburskiy (2):
ve sysfs: export sysfs_perms_set helper
module: export sysfs dentries in containers
fs/sysfs/ve.c |2 +-
include/linux/sysfs.h |5 +
kernel/module.c
time and there is not need to allocate this buffer each
time we need to expose or hide module sysfs dentries (allocating on stack
gives "the frame size of 4104 bytes is larger than 2048 bytes" warning)
https://jira.sw.ru/browse/PSBM-63892
Signed-off-by: Stanislav
It's going to be used to expose modules sysfs dentries
Signed-off-by: Stanislav Kinsburskiy
---
fs/sysfs/ve.c |2 +-
include/linux/sysfs.h |5 +
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/fs/sysfs/ve.c b/fs/sysfs/ve.c
index d7eff88..4c6f9ce 100644
---
I'm sorry, but this patch can and has to be split into series of patches.
11 июля 2017 г. 20:48 пользователь Dmitry Safonov
написал:
Mainstream already has arch_prctl(MAP_VDSO_64), but this was
ommited for simplicity and we only have arch_prctl(MAP_VDSO_32).
This was not a problem as previously
vars moved to ipvs struct.")
Signed-off-by: Tommi Rantala
Acked-by: Julian Anastasov
Signed-off-by: Simon Horman
Signed-off-by: Stanislav Kinsburskiy
---
net/netfilter/ipvs/ip_vs_ctl.c |3 +++
1 file changed, 3 insertions(+)
diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/
-by: Andrew Morton
Signed-off-by: Linus Torvalds
Signed-off-by: Stanislav Kinsburskiy
---
fs/proc/base.c |1 +
1 file changed, 1 insertion(+)
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 0af670c..291e503 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -2028,6 +2028,7 @@ static int proc_
low.c by using this latter form of callback
registration.
Cc: Li RongQing
Cc: Sasha Levin
Cc: Andrew Morton
Cc: Chris Metcalf
Cc: Ingo Molnar
Acked-by: David S. Miller
Signed-off-by: Srivatsa S. Bhat
Signed-off-by: Rafael J. Wysocki
Signed-off-by: Stanislav Kinsburskiy
---
net/core/flow.
Initial patch was supporting only slave mounts as bind-mounts.
https://jira.sw.ru/browse/PSBM-68663
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/criu/mount.c b/criu/mount.c
index 3abdb6e..7ed5d6e
Current check for unsupported mounts is obviously broken. So, skip it for a
while.
https://jira.sw.ru/browse/PSBM-68708
Signed-off-by: Stanislav Kinsburskiy
---
criu/mount.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/criu/mount.c b/criu/mount.c
index 7ed5d6e..2ec94b3 100644
--- a
19 июля 2017 г. 9:14 PM пользователь Andrey Vagin
написал:
On Wed, Jul 19, 2017 at 08:04:22PM +0300, Andrey Ryabinin wrote:
>
>
> On 07/19/2017 04:14 AM, Andrei Vagin wrote:
> > From: Andrei Vagin
> >
> > task_work_run() has to be called before exit_task_namespaces(),
> > because fuse_abort_co
19 июля 2017 г. 9:37 PM пользователь Andrey Vagin
написал:
On Wed, Jul 19, 2017 at 11:31:28AM -0700, Stanislav Kinsburskiy wrote:
>
>
> 19 июля 2017 г. 9:14 PM пользователь Andrey Vagin
> написал:
>
> On Wed, Jul 19, 2017 at 08:04:22PM +0300, And
Signed-off-by: Stanislav Kinsburskiy
---
include/linux/proc_fs.h |7 +++
1 file changed, 7 insertions(+)
diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h
index 4660327..1b516bc 100644
--- a/include/linux/proc_fs.h
+++ b/include/linux/proc_fs.h
@@ -67,6 +67,13 @@ static
Revert "proc-net: mark all files in proc-net directories as virtualized".
This tirck doesn't allow to define any non-visible per-net dentry, while we
will need it it to support container migration with disabled netfilter.
Signed-off-by: Stanislav Kinsburskiy
---
fs/proc/generic
This is precursor series in scope of https://jira.sw.ru/browse/PSBM-58574
---
Stanislav Kinsburskiy (3):
proc-net: revert e5990221bde54d757a5b5d85e81b2535064570ad
proc-net: proc_net_create() helper introduced
proc-net: virtualize all the network proc entries
drivers/net/ppp
Use proc_net_create() and proc_net_create_data() for all the per-net dentries.
We need this to expose per-net sysfs dentries once commit
e5990221bde54d757a5b5d85e81b2535064570ad is reverted
Signed-off-by: Stanislav Kinsburskiy
---
drivers/net/ppp/pppoe.c|2
1 - 100 of 430 matches
Mail list logo