Jan Engelhardt wrote:
On Oct 9 2007 09:26, Vasily Averin wrote:
On one of our servers timer interrupts (i.e irq0) are stops working. As
result
any kernel timers do not triggers and tasks waiting some signals from timers
hangs forever.
What kernel.. and tried CONFIG_NO_HZ=n?
Originally
Hi, Balbir-san
This is a patch set against memory cgroup I have now.
Reflected comments I got.
=
[1] charge refcnt fix patch - avoid charging against a page which is being
uncharged.
[2] fix-err-handling patch - remove unnecesary unlock_page_cgroup()
The logic of uncharging is
- decrement refcnt - lock page cgroup - remove page cgroup.
But the logic of charging is
- lock page cgroup - increment refcnt - return.
Then, one charge will be added to a page_cgroup under being removed.
This makes no big trouble (like panic) but one charge is
This unlock_page_cgroup() is unnecessary.
Signed-off-by: KAMEZAWA Hiroyuki [EMAIL PROTECTED]
mm/memcontrol.c |2 --
1 file changed, 2 deletions(-)
Index: linux-2.6.23-rc8-mm2/mm/memcontrol.c
===
---
This patch makes mem_cgroup_isolate_pages() to be
- ignore !PageLRU pages.
- fixes the bug that it makes no progress if page_zone(page) != zone
page once find. (just increment scan in this case.)
kswapd and memory migraion removes a page from list when it handles
a page for
While using memory control cgroup, page-migration under it works as following.
==
1. uncharge all refs at try to unmap.
2. charge regs again remove_migration_ptes()
==
This is simple but has following problems.
==
The page is uncharged and chaged back again if *mapped*.
- This means that
This patch adds an interface memory.force_reclaim.
Any write to this file will drop all charges in this cgroup if
there is no task under.
%echo 1 /../memory.force_reclaim
will drop all charges of memory cgroup if cgroup's tasks is empty.
This is useful to invoke rmdir() against memory
This patch adds follwoing functions.
- clear_page_cgroup(page, pc)
- page_cgroup_assign_new_page_group(page, pc)
Mainly for cleaunp.
A manner check page-cgroup again after lock_page_cgroup() is
implemented in straight way.
Signed-off-by: KAMEZAWA Hiroyuki [EMAIL PROTECTED]
On Tue, 9 Oct 2007, Vasily Averin wrote:
On one of our servers timer interrupts (i.e irq0) are stops working. As result
any kernel timers do not triggers and tasks waiting some signals from timers
hangs forever.
Which kernel version ?
Most noticeable effect of this situation is that any
KAMEZAWA Hiroyuki wrote:
Hi, Balbir-san
This is a patch set against memory cgroup I have now.
Reflected comments I got.
=
[1] charge refcnt fix patch - avoid charging against a page which is
being
uncharged.
[2] fix-err-handling patch -
KAMEZAWA Hiroyuki wrote:
The logic of uncharging is
- decrement refcnt - lock page cgroup - remove page cgroup.
But the logic of charging is
- lock page cgroup - increment refcnt - return.
Then, one charge will be added to a page_cgroup under being removed.
This makes no big trouble
KAMEZAWA Hiroyuki wrote:
This patch adds follwoing functions.
- clear_page_cgroup(page, pc)
- page_cgroup_assign_new_page_group(page, pc)
Mainly for cleaunp.
A manner check page-cgroup again after lock_page_cgroup() is
implemented in straight way.
Signed-off-by: KAMEZAWA
On Tue, 09 Oct 2007 16:39:48 +0530
Balbir Singh [EMAIL PROTECTED] wrote:
+static inline int
+page_cgroup_assign_new_page_cgroup(struct page *page, struct page_cgroup
*pc)
+{
+ int ret = 0;
+
+ lock_page_cgroup(page);
+ if (!page_get_page_cgroup(page))
+
Jeff Garzik wrote:
Pavel Emelyanov wrote:
The unregister_netdevice() and dev_change_net_namespace()
both check for dev-flags to be IFF_UP before calling the
dev_close(), but the dev_close() checks for IFF_UP itself,
so remove those unneeded checks.
Signed-off-by: Pavel Emelyanov [EMAIL
The newly created net namespace is set to 0 with memset()
in setup_net(). The setup_net() is also called for the
init_net_ns(), which is zeroed naturally as a global var.
So remove this memset and allocate new nets with the
kmem_cache_zalloc().
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
Currently indexes for netdevices come sequentially one by
one, and the same stays true even for devices that are
created for namespaces.
Side effects of this are:
* lo device has not 1 index in a namespace. This may break
some userspace that relies on it (and AFAIR something
really broke
Pavel Emelyanov wrote:
The unregister_netdevice() and dev_change_net_namespace()
both check for dev-flags to be IFF_UP before calling the
dev_close(), but the dev_close() checks for IFF_UP itself,
so remove those unneeded checks.
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
---
diff
KAMEZAWA Hiroyuki wrote:
This patch makes mem_cgroup_isolate_pages() to be
- ignore !PageLRU pages.
- fixes the bug that it makes no progress if page_zone(page) != zone
page once find. (just increment scan in this case.)
kswapd and memory migraion removes a page from list when it
Pavel Emelyanov wrote:
Currently indexes for netdevices come sequentially one by
one, and the same stays true even for devices that are
created for namespaces.
Side effects of this are:
* lo device has not 1 index in a namespace. This may break
some userspace that relies on it (and AFAIR
Many (very many) seq files in net/ allocate some private data
to use it later (mostly for iteration state). All this code
was obviously get using copy-paste method, so move it into
one place.
Almost all of these places either set this private to 0, or
keep uninitialized. Some places, however,
This function allocates the zeroed chunk of memory and
call seq_open(). The __seq_open_private() helper returns
the allocated memory to make it possible for the caller
to initialize it.
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
---
diff --git a/fs/seq_file.c b/fs/seq_file.c
index
This concerns the ipv4 and ipv6 code mostly, but also the netlink
and unix sockets.
The netlink code is an example of how to use the __seq_open_private()
call - it saves the net namespace on this private.
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
---
diff --git a/net/ipv4/arp.c
Just switch to the consolidated code.
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
Cc: Patrick Caulfield [EMAIL PROTECTED]
---
diff --git a/net/decnet/dn_neigh.c b/net/decnet/dn_neigh.c
index b66e3be..e851b14 100644
--- a/net/decnet/dn_neigh.c
+++ b/net/decnet/dn_neigh.c
@@ -580,24 +580,8
Just switch to the consolidated code
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
Cc: Samuel Ortiz [EMAIL PROTECTED]
---
diff --git a/net/irda/irlap.c b/net/irda/irlap.c
index 3d76aaf..f3236ac 100644
--- a/net/irda/irlap.c
+++ b/net/irda/irlap.c
@@ -1219,29 +1219,11 @@ static const struct
Just switch to the consolidated code.
Signed-off-by: Pavel Emelyanov [EMAIL PROTECTED]
Cc: Neil Brown [EMAIL PROTECTED]
---
diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c
index ebe344f..8e05557 100644
--- a/net/sunrpc/cache.c
+++ b/net/sunrpc/cache.c
@@ -1218,23 +1218,15 @@ static const
KAMEZAWA Hiroyuki wrote:
While using memory control cgroup, page-migration under it works as following.
==
1. uncharge all refs at try to unmap.
2. charge regs again remove_migration_ptes()
==
This is simple but has following problems.
==
The page is uncharged and chaged back again if
This patch add a pernet subsystem to register /proc/net/udplite
per network namespace.
Signed-off-by: Daniel Lezcano [EMAIL PROTECTED]
---
net/ipv4/udplite.c | 24 ++--
1 file changed, 22 insertions(+), 2 deletions(-)
Index: linux-2.6-netns/net/ipv4/udplite.c
The following patches make the udplite protocol per network namespace.
They are only a couple of patches because the core udp engine per
network namespace is already done and udplite relies on it.
--
___
Containers mailing list
[EMAIL PROTECTED]
This patch activates the udplite protocol for multiple
network namespaces.
Signed-off-by: Daniel Lezcano [EMAIL PROTECTED]
---
net/ipv4/udplite.c | 12 +---
1 file changed, 1 insertion(+), 11 deletions(-)
Index: linux-2.6-netns/net/ipv4/udplite.c
Pavel Emelyanov [EMAIL PROTECTED] writes:
Currently indexes for netdevices come sequentially one by
one, and the same stays true even for devices that are
created for namespaces.
Side effects of this are:
* lo device has not 1 index in a namespace. This may break
some userspace that
David Stevens [EMAIL PROTECTED] writes:
Sorry if this is a dumb question, but what is the model you intend for
SNMP? Do you want each namespace to be its own virtual machine with
its own, separate MIB?
Each network namespace appears to user space as a completely separate
network stack. So
KAMEZAWA Hiroyuki wrote:
This patch adds an interface memory.force_reclaim.
Any write to this file will drop all charges in this cgroup if
there is no task under.
%echo 1 /../memory.force_reclaim
Looks like a good name, do you think system administrators would
find force_empty more
From: Pavel Emelyanov [EMAIL PROTECTED]
Date: Tue, 09 Oct 2007 15:48:19 +0400
The newly created net namespace is set to 0 with memset()
in setup_net(). The setup_net() is also called for the
init_net_ns(), which is zeroed naturally as a global var.
So remove this memset and allocate new
From: [EMAIL PROTECTED] (Eric W. Biederman)
Date: Tue, 09 Oct 2007 11:43:58 -0600
David Stevens [EMAIL PROTECTED] writes:
Sorry if this is a dumb question, but what is the model you intend for
SNMP? Do you want each namespace to be its own virtual machine with
its own, separate MIB?
From: David Stevens [EMAIL PROTECTED]
Date: Tue, 9 Oct 2007 09:18:25 -0700
Ifindex's have to uniquely identify the interface (virtual or
otherwise) to remote queriers (not just local applications), so
unless you pay the price of separating all the SNMP MIBs per
namespace too, it seems you'll
From 945fe66259cd0cfdc2fe846287b7821e329a558c Mon Sep 17 00:00:00 2001
From: [EMAIL PROTECTED] [EMAIL PROTECTED](none)
Date: Tue, 9 Oct 2007 08:30:30 -0700
Subject: [PATCH] namespaces: introduce sys_hijack (v4)
Move most of do_fork() into a new do_fork_task() which acts on
a new argument, task,
From: Pavel Emelyanov [EMAIL PROTECTED]
Date: Tue, 09 Oct 2007 16:19:25 +0400
Currently indexes for netdevices come sequentially one by
one, and the same stays true even for devices that are
created for namespaces.
Side effects of this are:
* lo device has not 1 index in a namespace.
David Miller [EMAIL PROTECTED] writes:
From: [EMAIL PROTECTED] (Eric W. Biederman)
Date: Tue, 09 Oct 2007 11:43:58 -0600
David Stevens [EMAIL PROTECTED] writes:
Sorry if this is a dumb question, but what is the model you intend for
SNMP? Do you want each namespace to be its own virtual
From: [EMAIL PROTECTED] (Eric W. Biederman)
Date: Tue, 09 Oct 2007 15:00:10 -0600
Regardless it is early yet and there is plenty of time to revisit this
after we solved the easier and less controversial problems.
Ok.
I would encourage you to learn how the SNMP mibs work, and whether
they
The logic of uncharging is
- decrement refcnt - lock page cgroup - remove page cgroup.
But the logic of charging is
- lock page cgroup - increment refcnt - return.
Then, one charge will be added to a page_cgroup under being removed.
This makes no big trouble (like panic) but one charge
On Fri, Oct 05, 2007 at 06:44:04AM -0600, Eric W. Biederman wrote:
Greg KH [EMAIL PROTECTED] writes:
I would be interested in seeing what your patches look like.
Sure.
I don't
think that we should take any more sysfs changes for 2.6.24 as we do
have a lot of them right now, and I
On Fri, Oct 05, 2007 at 06:12:41AM -0600, Eric W. Biederman wrote:
Greg KH [EMAIL PROTECTED] writes:
Also fun is that the dev file implementation needs to be able to
report different major:minor numbers based on which mount of
sysfs we are dealing with.
Um, no, that's not going
On Wed, 10 Oct 2007 07:31:38 +0900 (JST)
[EMAIL PROTECTED] (YAMAMOTO Takashi) wrote:
- atomic_inc(pc-ref_cnt);
- goto done;
+ if (unlikely(!atomic_inc_not_zero(pc-ref_cnt))) {
+ /* this page is under being uncharge ? */
+
On Tue, 9 Oct 2007 20:26:42 +0900
KAMEZAWA Hiroyuki [EMAIL PROTECTED] wrote:
+ */
+ if (clear_page_cgroup(page, pc) == pc) {
OK.. so we've come so far and seen that pc has changed underneath us,
what do we do with this pc?
Hmm... How about this ?
==
if
hi,
i implemented some statistics for your memory controller.
here's a new version.
changes from the previous:
- make counters per-cpu.
- value *= PAGE_SIZE
YAMAMOTO Takashi
--- linux-2.6.23-rc8-mm2-stat/mm/memcontrol.c.BACKUP2007-10-01
17:19:57.0 +0900
+++
Jeff Garzik [EMAIL PROTECTED] wrote:
One side effect of this patch: might_sleep() is now called unconditionally.
That would be seem to be a good thing :)
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmVHI~} [EMAIL PROTECTED]
Home Page: http://gondor.apana.org.au/~herbert/
46 matches
Mail list logo
