Re: [PATCH -mm 1/3] sysv ipc: increase msgmnb default value wrt.the number of cpus
From: Nadia Derbey
Date: Wed Jun 25 2008 - 06:34:47 EST
Andrew Morton wrote:
On Tue, 24 Jun 2008 11:34:53 +0200
<Solofo.Ramangalahy@xxxxxxxx> wrote:
From: Solofo Ramangalahy <Solofo.Ramangalahy@xxxxxxxx>
Initialize msgmnb value to
min(MSGMNB * num_online_cpus(), MSGMNB * MSG_CPU_SCALE)
to increase the default value for larger machines.
MSG_CPU_SCALE scaling factor is defined to be 4, as 16384 x 4 = 65536
is an already used and recommended value.
The msgmni value is made dependant of msgmnb to keep the memory
dedicated to message queues within the 1/MSG_MEM_SCALE of lowmem
bound.
Unlike msgmni, the value is not scaled (down) with respect to the
number of ipc namespaces for simplicity.
To disable recomputation when user explicitely set a value,
we reuse the callback defined for msgmni.
As msgmni and msgmnb are correlated, user settings of any of the two
disable recomputation of both, for now. This is refined in a later
patch.
When a negative value is put in /proc/sys/kernel/msgmnb
automatic recomputing is re-enabled.
Thanks for taking the time to describe this work so well.
---
Documentation/sysctl/kernel.txt | 28 ++++++++++++++++++++++++++++
include/linux/msg.h | 6 ++++++
ipc/ipc_sysctl.c | 5 +++--
ipc/msg.c | 17 +++++++++++++----
4 files changed, 50 insertions(+), 6 deletions(-)
Index: b/ipc/msg.c
===================================================================
--- a/ipc/msg.c
+++ b/ipc/msg.c
@@ -38,6 +38,7 @@
#include <linux/rwsem.h>
#include <linux/nsproxy.h>
#include <linux/ipc_namespace.h>
+#include <linux/cpumask.h>
#include <asm/current.h>
#include <asm/uaccess.h>
@@ -92,7 +93,7 @@ void recompute_msgmni(struct ipc_namespa
si_meminfo(&i);
allowed = (((i.totalram - i.totalhigh) / MSG_MEM_SCALE) * i.mem_unit)
- / MSGMNB;
+ / ns->msg_ctlmnb;
nb_ns = atomic_read(&nr_ipc_ns);
allowed /= nb_ns;
@@ -108,11 +109,19 @@ void recompute_msgmni(struct ipc_namespa
ns->msg_ctlmni = allowed;
}
+/*
+ * Scale msgmnb with the number of online cpus, up to 4x MSGMNB.
+ */
+void recompute_msgmnb(struct ipc_namespace *ns)
+{
+ ns->msg_ctlmnb =
+ min(MSGMNB * num_online_cpus(), MSGMNB * MSG_CPU_SCALE);
+}
void msg_init_ns(struct ipc_namespace *ns)
{
ns->msg_ctlmax = MSGMAX;
- ns->msg_ctlmnb = MSGMNB;
+ recompute_msgmnb(ns);
recompute_msgmni(ns);
@@ -132,8 +141,8 @@ void __init msg_init(void)
{
msg_init_ns(&init_ipc_ns);
- printk(KERN_INFO "msgmni has been set to %d\n",
- init_ipc_ns.msg_ctlmni);
+ printk(KERN_INFO "msgmni has been set to %d, msgmnb to %d\n",
+ init_ipc_ns.msg_ctlmni, init_ipc_ns.msg_ctlmnb);
ipc_init_proc_interface("sysvipc/msg",
" key msqid perms cbytes qnum lspid lrpid uid gid cuid cgid stime rtime ctime\n",
Index: b/include/linux/msg.h
===================================================================
--- a/include/linux/msg.h
+++ b/include/linux/msg.h
@@ -58,6 +58,12 @@ struct msginfo {
* more than 16 GB : msgmni = 32K (IPCMNI)
*/
#define MSG_MEM_SCALE 32
+/*
+ * Scaling factor to compute msgmnb: ns->msg_ctlmnb is between MSGMNB
+ * and MSGMNB * MSG_CPU_SCALE. This leads to a max msgmnb value of
+ * 65536 which is an already used and recommended value.
+ */
+#define MSG_CPU_SCALE 4
#define MSGMNI 16 /* <= IPCMNI */ /* max # of msg queue identifiers */
#define MSGMAX 8192 /* <= INT_MAX */ /* max size of message (bytes) */
Index: b/ipc/ipc_sysctl.c
===================================================================
--- a/ipc/ipc_sysctl.c
+++ b/ipc/ipc_sysctl.c
@@ -42,6 +42,7 @@ static void tunable_set_callback(int val
* Re-enable automatic recomputing only if not already
* enabled.
*/
+ recompute_msgmnb(current->nsproxy->ipc_ns);
recompute_msgmni(current->nsproxy->ipc_ns);
cond_register_ipcns_notifier(current->nsproxy->ipc_ns);
}
@@ -210,8 +211,8 @@ static struct ctl_table ipc_kern_table[]
.data = &init_ipc_ns.msg_ctlmnb,
.maxlen = sizeof (init_ipc_ns.msg_ctlmnb),
.mode = 0644,
- .proc_handler = proc_ipc_dointvec,
- .strategy = sysctl_ipc_data,
+ .proc_handler = proc_ipc_callback_dointvec,
+ .strategy = sysctl_ipc_registered_data,
},
{
.ctl_name = KERN_SEM,
Index: b/Documentation/sysctl/kernel.txt
===================================================================
--- a/Documentation/sysctl/kernel.txt
+++ b/Documentation/sysctl/kernel.txt
@@ -179,6 +179,34 @@ kernel stack.
==============================================================
+msgmnb
+
+Maximum size in bytes (not in message count) of a single SystemV IPC
+message queue (b stands for bytes).
+
+This value is dynamic and depends on the online cpu count of the
+machine (taking cpu hotplug into account).
+
+Computed values are between MSGMNB and MSGMNB*MSG_CPU_SCALE #define
+constants (currently [16384,65536]).
+
+The exact value is automatically (re)computed, but:
+. If the value is positioned from user space (via procfs or sysctl()),
+ to a positive value then the automatic recomputation is
+ disabled. This leaves control to user space. E.g.
+
+ # echo 16384 > /proc/sys/kernel/msgmnb
+
+. If the value is positioned from user space to a negative value, then
+ the computation is reenabled. E.g.
+
+ # echo -1 > /proc/sys/kernel/msgmnb
+
+See recompute_msgmnb() function in ipc/ directory for details.
+The value of msgmnb is coupled with the value of msgmni.
+
The magical positive-versus-negative number trick is a bit obscure, and
I don't think there's any precedent for it in the kernel ABI (which is
what this is).
Is there anything we can do to reduce the unusualness of this
interface? Say, add a new /proc/sys/kernel/automatic-msgmnb which
contains the automatic scaling and leave /proc/sys/kernel/msgmnb
containing the manual scaling? Or something like that?
Well, I plead guilty ;-)
I've done this proposal when sending the msgmni scaling patches
(unfortunatly my network is down, so can't look the reference thread).
From what I have in my folders here's the complete story:
. January 08: sent the patches
. 02/05/2008: got an answer from Yasunori Goto:
Yasunori Goto wrote:
> Hmmm. I suppose this may be side effect which user does not wish.
>
> I would like to recommend there should be a switch which can turn
> on/off
> automatic recomputing.
> If user would like to change this value, it should be turned off.
> Otherwise, his requrest will be rejected with some messages.
>
> Probably, user can understand easier than this side effect.
. 02/11/2008: resent the patches after fixing the issues:
Nadia Derbey wrote:
> Resending the set of patches after Yasunori's remark about being able
> to turn on/off automatic recomputing.
> (see message at http://lkml.org/lkml/2008/2/5/149).
> I actually introduced an intermediate solution: when msgmni is set by
> hand, it is uneregistered from the ipcns notifier chain (i.e.
> automatic recomputing is disabled). This corresponds to an implicit
> turn off. Setting it to a negative value makes it registered back in
> the notifier chain (which corresponds to the turn on proposed by
> Yasunaori).
And I don't remember anybody complaining about that :-(
Sorry for introducing this "magical positive-vs-negative # trick".
Will think a bit more about your suggestion.
Regards,
Nadia
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/