2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* linux/ipc/msg.c
|
2006-07-30 18:04:11 +08:00
|
|
|
* Copyright (C) 1992 Krishna Balasubramanian
|
2005-04-17 06:20:36 +08:00
|
|
|
*
|
|
|
|
* Removed all the remaining kerneld mess
|
|
|
|
* Catch the -EFAULT stuff properly
|
|
|
|
* Use GFP_KERNEL for messages as in 1.2
|
|
|
|
* Fixed up the unchecked user space derefs
|
|
|
|
* Copyright (C) 1998 Alan Cox & Andi Kleen
|
|
|
|
*
|
|
|
|
* /proc/sysvipc/msg support (c) 1999 Dragos Acostachioaie <dragos@iname.com>
|
|
|
|
*
|
|
|
|
* mostly rewritten, threaded and wake-one semantics added
|
|
|
|
* MSGMAX limit removed, sysctl's added
|
2006-01-15 09:43:54 +08:00
|
|
|
* (c) 1999 Manfred Spraul <manfred@colorfullife.com>
|
2006-04-03 05:07:33 +08:00
|
|
|
*
|
|
|
|
* support for audit of ipc object properties and permission changes
|
|
|
|
* Dustin Kirkland <dustin.kirkland@us.ibm.com>
|
2006-10-02 17:18:21 +08:00
|
|
|
*
|
|
|
|
* namespaces support
|
|
|
|
* OpenVZ, SWsoft Inc.
|
|
|
|
* Pavel Emelianov <xemul@openvz.org>
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
|
2006-01-12 04:17:46 +08:00
|
|
|
#include <linux/capability.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/msg.h>
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
#include <linux/init.h>
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
#include <linux/mm.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/proc_fs.h>
|
|
|
|
#include <linux/list.h>
|
|
|
|
#include <linux/security.h>
|
|
|
|
#include <linux/sched.h>
|
|
|
|
#include <linux/syscalls.h>
|
|
|
|
#include <linux/audit.h>
|
2005-09-07 06:17:10 +08:00
|
|
|
#include <linux/seq_file.h>
|
2007-10-19 14:40:54 +08:00
|
|
|
#include <linux/rwsem.h>
|
2006-10-02 17:18:21 +08:00
|
|
|
#include <linux/nsproxy.h>
|
2008-02-08 20:18:22 +08:00
|
|
|
#include <linux/ipc_namespace.h>
|
2006-03-26 17:37:17 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <asm/current.h>
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
#include "util.h"
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
/*
|
|
|
|
* one msg_receiver structure for each sleeping receiver:
|
|
|
|
*/
|
2005-04-17 06:20:36 +08:00
|
|
|
struct msg_receiver {
|
2006-07-30 18:04:11 +08:00
|
|
|
struct list_head r_list;
|
|
|
|
struct task_struct *r_tsk;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
int r_mode;
|
|
|
|
long r_msgtype;
|
|
|
|
long r_maxsize;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-11-05 01:55:00 +08:00
|
|
|
struct msg_msg *volatile r_msg;
|
2005-04-17 06:20:36 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
/* one msg_sender for each sleeping sender */
|
|
|
|
struct msg_sender {
|
2006-07-30 18:04:11 +08:00
|
|
|
struct list_head list;
|
|
|
|
struct task_struct *tsk;
|
2005-04-17 06:20:36 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
#define SEARCH_ANY 1
|
|
|
|
#define SEARCH_EQUAL 2
|
|
|
|
#define SEARCH_NOTEQUAL 3
|
|
|
|
#define SEARCH_LESSEQUAL 4
|
2013-05-01 10:14:54 +08:00
|
|
|
#define SEARCH_NUMBER 5
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-02-08 20:18:57 +08:00
|
|
|
#define msg_ids(ns) ((ns)->ids[IPC_MSG_IDS])
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-10-02 17:18:21 +08:00
|
|
|
#define msg_unlock(msq) ipc_unlock(&(msq)->q_perm)
|
|
|
|
|
2008-02-08 20:18:57 +08:00
|
|
|
static void freeque(struct ipc_namespace *, struct kern_ipc_perm *);
|
2007-10-19 14:40:49 +08:00
|
|
|
static int newque(struct ipc_namespace *, struct ipc_params *);
|
2005-04-17 06:20:36 +08:00
|
|
|
#ifdef CONFIG_PROC_FS
|
2005-09-07 06:17:10 +08:00
|
|
|
static int sysvipc_msg_proc_show(struct seq_file *s, void *it);
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif
|
|
|
|
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
/*
|
|
|
|
* Scale msgmni with the available lowmem size: the memory dedicated to msg
|
|
|
|
* queues should occupy at most 1/MSG_MEM_SCALE of lowmem.
|
2008-04-29 16:00:40 +08:00
|
|
|
* Also take into account the number of nsproxies created so far.
|
|
|
|
* This should be done staying within the (MSGMNI , IPCMNI/nr_ipc_ns) range.
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
*/
|
2008-04-29 16:00:42 +08:00
|
|
|
void recompute_msgmni(struct ipc_namespace *ns)
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
{
|
|
|
|
struct sysinfo i;
|
|
|
|
unsigned long allowed;
|
2008-04-29 16:00:40 +08:00
|
|
|
int nb_ns;
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
|
|
|
|
si_meminfo(&i);
|
|
|
|
allowed = (((i.totalram - i.totalhigh) / MSG_MEM_SCALE) * i.mem_unit)
|
|
|
|
/ MSGMNB;
|
2008-04-29 16:00:40 +08:00
|
|
|
nb_ns = atomic_read(&nr_ipc_ns);
|
|
|
|
allowed /= nb_ns;
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
|
|
|
|
if (allowed < MSGMNI) {
|
|
|
|
ns->msg_ctlmni = MSGMNI;
|
2008-06-06 13:46:38 +08:00
|
|
|
return;
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
}
|
|
|
|
|
2008-04-29 16:00:40 +08:00
|
|
|
if (allowed > IPCMNI / nb_ns) {
|
|
|
|
ns->msg_ctlmni = IPCMNI / nb_ns;
|
2008-06-06 13:46:38 +08:00
|
|
|
return;
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
ns->msg_ctlmni = allowed;
|
|
|
|
}
|
|
|
|
|
2008-02-08 20:18:57 +08:00
|
|
|
void msg_init_ns(struct ipc_namespace *ns)
|
2006-10-02 17:18:21 +08:00
|
|
|
{
|
|
|
|
ns->msg_ctlmax = MSGMAX;
|
|
|
|
ns->msg_ctlmnb = MSGMNB;
|
ipc: scale msgmni to the amount of lowmem
On large systems we'd like to allow a larger number of message queues. In
some cases up to 32K. However simply setting MSGMNI to a larger value may
cause problems for smaller systems.
The first patch of this series introduces a default maximum number of message
queue ids that scales with the amount of lowmem.
Since msgmni is per namespace and there is no amount of memory dedicated to
each namespace so far, the second patch of this series scales msgmni to the
number of ipc namespaces too.
Since msgmni depends on the amount of memory, it becomes necessary to
recompute it upon memory add/remove. In the 4th patch, memory hotplug
management is added: a notifier block is registered into the memory hotplug
notifier chain for the ipc subsystem. Since the ipc namespaces are not linked
together, they have their own notification chain: one notifier_block is
defined per ipc namespace. Each time an ipc namespace is created (removed) it
registers (unregisters) its notifier block in (from) the ipcns chain. The
callback routine registered in the memory chain invokes the ipcns notifier
chain with the IPCNS_MEMCHANGE event. Each callback routine registered in the
ipcns namespace, in turn, recomputes msgmni for the owning namespace.
The 5th patch makes it possible to keep the memory hotplug notifier chain's
lock for a lesser amount of time: instead of directly notifying the ipcns
notifier chain upon memory add/remove, a work item is added to the global
workqueue. When activated, this work item is the one who notifies the ipcns
notifier chain.
Since msgmni depends on the number of ipc namespaces, it becomes necessary to
recompute it upon ipc namespace creation / removal. The 6th patch uses the
ipc namespace notifier chain for that purpose: that chain is notified each
time an ipc namespace is created or removed. This makes it possible to
recompute msgmni for all the namespaces each time one of them is created or
removed.
When msgmni is explicitely set from userspace, we should avoid recomputing it
upon memory add/remove or ipcns creation/removal. This is what the 7th patch
does: it simply unregisters the ipcns callback routine as soon as msgmni has
been changed from procfs or sysctl().
Even if msgmni is set by hand, it should be possible to make it back
automatically recomputed upon memory add/remove or ipcns creation/removal.
This what is achieved in patch 8: if set to a negative value, msgmni is added
back to the ipcns notifier chain, making it automatically recomputed again.
This patch:
Compute msg_ctlmni to make it scale with the amount of lowmem. msg_ctlmni is
now set to make the message queues occupy 1/32 of the available lowmem.
Some cleaning has also been done for the MSGPOOL constant: the msgctl man page
says it's not used, but it also defines it as a size in bytes (the code
expresses it in Kbytes).
Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Pierre Peiffer <pierre.peiffer@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 16:00:39 +08:00
|
|
|
|
|
|
|
recompute_msgmni(ns);
|
|
|
|
|
2007-10-19 14:40:56 +08:00
|
|
|
atomic_set(&ns->msg_bytes, 0);
|
|
|
|
atomic_set(&ns->msg_hdrs, 0);
|
2008-02-08 20:18:57 +08:00
|
|
|
ipc_init_ids(&ns->ids[IPC_MSG_IDS]);
|
2006-10-02 17:18:21 +08:00
|
|
|
}
|
|
|
|
|
2008-02-08 20:18:22 +08:00
|
|
|
#ifdef CONFIG_IPC_NS
|
2006-10-02 17:18:21 +08:00
|
|
|
void msg_exit_ns(struct ipc_namespace *ns)
|
|
|
|
{
|
2008-02-08 20:18:57 +08:00
|
|
|
free_ipcs(ns, &msg_ids(ns), freeque);
|
2009-12-16 08:47:27 +08:00
|
|
|
idr_destroy(&ns->ids[IPC_MSG_IDS].ipcs_idr);
|
2006-10-02 17:18:21 +08:00
|
|
|
}
|
2008-02-08 20:18:22 +08:00
|
|
|
#endif
|
2006-10-02 17:18:21 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
void __init msg_init(void)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2008-02-08 20:18:57 +08:00
|
|
|
msg_init_ns(&init_ipc_ns);
|
2008-06-06 13:46:38 +08:00
|
|
|
|
|
|
|
printk(KERN_INFO "msgmni has been set to %d\n",
|
|
|
|
init_ipc_ns.msg_ctlmni);
|
|
|
|
|
2005-09-07 06:17:10 +08:00
|
|
|
ipc_init_proc_interface("sysvipc/msg",
|
|
|
|
" key msqid perms cbytes qnum lspid lrpid uid gid cuid cgid stime rtime ctime\n",
|
2006-10-02 17:18:21 +08:00
|
|
|
IPC_MSG_IDS, sysvipc_msg_proc_show);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2007-10-19 14:40:54 +08:00
|
|
|
/*
|
|
|
|
* msg_lock_(check_) routines are called in the paths where the rw_mutex
|
|
|
|
* is not held.
|
|
|
|
*/
|
2007-10-19 14:40:51 +08:00
|
|
|
static inline struct msg_queue *msg_lock(struct ipc_namespace *ns, int id)
|
|
|
|
{
|
2007-10-19 14:40:51 +08:00
|
|
|
struct kern_ipc_perm *ipcp = ipc_lock(&msg_ids(ns), id);
|
|
|
|
|
IPC: fix error check in all new xxx_lock() and xxx_exit_ns() functions
In the new implementation of the [sem|shm|msg]_lock[_check]() routines, we
use the return value of ipc_lock() in container_of() without any check.
But ipc_lock may return a errcode. The use of this errcode in
container_of() may alter this errcode, and we don't want this.
And in xxx_exit_ns, the pointer return by idr_find is of type 'struct
kern_ipc_per'...
Today, the code will work as is because the member used in these
container_of() is the first member of its container (offset == 0), the
errcode isn't changed then. But in the general case, we can't count on
this assumption and this may lead later to a real bug if we don't correct
this.
Again, the proposed solution is simple and correct. But, as pointed by
Nadia, with this solution, the same check will be done several times (in
all sub-callers...), what is not very funny/optimal...
Signed-off-by: Pierre Peiffer <pierre.peiffer@bull.net>
Cc: Nadia Derbey <Nadia.Derbey@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-06 17:36:23 +08:00
|
|
|
if (IS_ERR(ipcp))
|
|
|
|
return (struct msg_queue *)ipcp;
|
|
|
|
|
2007-10-19 14:40:51 +08:00
|
|
|
return container_of(ipcp, struct msg_queue, q_perm);
|
2007-10-19 14:40:51 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct msg_queue *msg_lock_check(struct ipc_namespace *ns,
|
|
|
|
int id)
|
|
|
|
{
|
2007-10-19 14:40:51 +08:00
|
|
|
struct kern_ipc_perm *ipcp = ipc_lock_check(&msg_ids(ns), id);
|
|
|
|
|
IPC: fix error check in all new xxx_lock() and xxx_exit_ns() functions
In the new implementation of the [sem|shm|msg]_lock[_check]() routines, we
use the return value of ipc_lock() in container_of() without any check.
But ipc_lock may return a errcode. The use of this errcode in
container_of() may alter this errcode, and we don't want this.
And in xxx_exit_ns, the pointer return by idr_find is of type 'struct
kern_ipc_per'...
Today, the code will work as is because the member used in these
container_of() is the first member of its container (offset == 0), the
errcode isn't changed then. But in the general case, we can't count on
this assumption and this may lead later to a real bug if we don't correct
this.
Again, the proposed solution is simple and correct. But, as pointed by
Nadia, with this solution, the same check will be done several times (in
all sub-callers...), what is not very funny/optimal...
Signed-off-by: Pierre Peiffer <pierre.peiffer@bull.net>
Cc: Nadia Derbey <Nadia.Derbey@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-06 17:36:23 +08:00
|
|
|
if (IS_ERR(ipcp))
|
|
|
|
return (struct msg_queue *)ipcp;
|
|
|
|
|
2007-10-19 14:40:51 +08:00
|
|
|
return container_of(ipcp, struct msg_queue, q_perm);
|
2007-10-19 14:40:51 +08:00
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:15 +08:00
|
|
|
static inline struct msg_queue *msq_obtain_object(struct ipc_namespace *ns, int id)
|
|
|
|
{
|
|
|
|
struct kern_ipc_perm *ipcp = ipc_obtain_object(&msg_ids(ns), id);
|
|
|
|
|
|
|
|
if (IS_ERR(ipcp))
|
|
|
|
return ERR_CAST(ipcp);
|
|
|
|
|
|
|
|
return container_of(ipcp, struct msg_queue, q_perm);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct msg_queue *msq_obtain_object_check(struct ipc_namespace *ns,
|
|
|
|
int id)
|
|
|
|
{
|
|
|
|
struct kern_ipc_perm *ipcp = ipc_obtain_object_check(&msg_ids(ns), id);
|
|
|
|
|
|
|
|
if (IS_ERR(ipcp))
|
|
|
|
return ERR_CAST(ipcp);
|
|
|
|
|
|
|
|
return container_of(ipcp, struct msg_queue, q_perm);
|
|
|
|
}
|
|
|
|
|
2007-10-19 14:40:48 +08:00
|
|
|
static inline void msg_rmid(struct ipc_namespace *ns, struct msg_queue *s)
|
|
|
|
{
|
|
|
|
ipc_rmid(&msg_ids(ns), &s->q_perm);
|
|
|
|
}
|
|
|
|
|
2007-10-19 14:40:53 +08:00
|
|
|
/**
|
|
|
|
* newque - Create a new msg queue
|
|
|
|
* @ns: namespace
|
|
|
|
* @params: ptr to the structure that contains the key and msgflg
|
|
|
|
*
|
2007-10-19 14:40:54 +08:00
|
|
|
* Called with msg_ids.rw_mutex held (writer)
|
2007-10-19 14:40:53 +08:00
|
|
|
*/
|
2007-10-19 14:40:49 +08:00
|
|
|
static int newque(struct ipc_namespace *ns, struct ipc_params *params)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct msg_queue *msq;
|
2006-07-30 18:04:11 +08:00
|
|
|
int id, retval;
|
2007-10-19 14:40:49 +08:00
|
|
|
key_t key = params->key;
|
|
|
|
int msgflg = params->flg;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
msq = ipc_rcu_alloc(sizeof(*msq));
|
|
|
|
if (!msq)
|
2005-04-17 06:20:36 +08:00
|
|
|
return -ENOMEM;
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
msq->q_perm.mode = msgflg & S_IRWXUGO;
|
2005-04-17 06:20:36 +08:00
|
|
|
msq->q_perm.key = key;
|
|
|
|
|
|
|
|
msq->q_perm.security = NULL;
|
|
|
|
retval = security_msg_queue_alloc(msq);
|
|
|
|
if (retval) {
|
|
|
|
ipc_rcu_putref(msq);
|
|
|
|
return retval;
|
|
|
|
}
|
|
|
|
|
ipc: move rcu lock out of ipc_addid
This patchset continues the work that began in the sysv ipc semaphore
scaling series, see
https://lkml.org/lkml/2013/3/20/546
Just like semaphores used to be, sysv shared memory and msg queues also
abuse the ipc lock, unnecessarily holding it for operations such as
permission and security checks.
This patchset mostly deals with mqueues, and while shared mem can be
done in a very similar way, I want to get these patches out in the open
first. It also does some pending cleanups, mostly focused on the two
level locking we have in ipc code, taking care of ipc_addid() and
ipcctl_pre_down_nolock() - yes there are still functions that need to be
updated as well.
This patch:
Make all callers explicitly take and release the RCU read lock.
This addresses the two level locking seen in newary(), newseg() and
newqueue(). For the last two, explicitly unlock the ipc object and the
rcu lock, instead of calling the custom shm_unlock and msg_unlock
functions. The next patch will deal with the open coded locking for
->perm.lock
Signed-off-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-09 07:01:09 +08:00
|
|
|
/* ipc_addid() locks msq upon success. */
|
2006-10-02 17:18:21 +08:00
|
|
|
id = ipc_addid(&msg_ids(ns), &msq->q_perm, ns->msg_ctlmni);
|
2007-10-19 14:40:57 +08:00
|
|
|
if (id < 0) {
|
2005-04-17 06:20:36 +08:00
|
|
|
security_msg_queue_free(msq);
|
|
|
|
ipc_rcu_putref(msq);
|
2007-10-19 14:40:57 +08:00
|
|
|
return id;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
msq->q_stime = msq->q_rtime = 0;
|
|
|
|
msq->q_ctime = get_seconds();
|
|
|
|
msq->q_cbytes = msq->q_qnum = 0;
|
2006-10-02 17:18:21 +08:00
|
|
|
msq->q_qbytes = ns->msg_ctlmnb;
|
2005-04-17 06:20:36 +08:00
|
|
|
msq->q_lspid = msq->q_lrpid = 0;
|
|
|
|
INIT_LIST_HEAD(&msq->q_messages);
|
|
|
|
INIT_LIST_HEAD(&msq->q_receivers);
|
|
|
|
INIT_LIST_HEAD(&msq->q_senders);
|
2007-10-19 14:40:48 +08:00
|
|
|
|
2013-07-09 07:01:11 +08:00
|
|
|
ipc_unlock_object(&msq->q_perm);
|
ipc: move rcu lock out of ipc_addid
This patchset continues the work that began in the sysv ipc semaphore
scaling series, see
https://lkml.org/lkml/2013/3/20/546
Just like semaphores used to be, sysv shared memory and msg queues also
abuse the ipc lock, unnecessarily holding it for operations such as
permission and security checks.
This patchset mostly deals with mqueues, and while shared mem can be
done in a very similar way, I want to get these patches out in the open
first. It also does some pending cleanups, mostly focused on the two
level locking we have in ipc code, taking care of ipc_addid() and
ipcctl_pre_down_nolock() - yes there are still functions that need to be
updated as well.
This patch:
Make all callers explicitly take and release the RCU read lock.
This addresses the two level locking seen in newary(), newseg() and
newqueue(). For the last two, explicitly unlock the ipc object and the
rcu lock, instead of calling the custom shm_unlock and msg_unlock
functions. The next patch will deal with the open coded locking for
->perm.lock
Signed-off-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-09 07:01:09 +08:00
|
|
|
rcu_read_unlock();
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-10-19 14:40:48 +08:00
|
|
|
return msq->q_perm.id;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static inline void ss_add(struct msg_queue *msq, struct msg_sender *mss)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2006-07-30 18:04:11 +08:00
|
|
|
mss->tsk = current;
|
|
|
|
current->state = TASK_INTERRUPTIBLE;
|
|
|
|
list_add_tail(&mss->list, &msq->q_senders);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static inline void ss_del(struct msg_sender *mss)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2006-07-30 18:04:11 +08:00
|
|
|
if (mss->list.next != NULL)
|
2005-04-17 06:20:36 +08:00
|
|
|
list_del(&mss->list);
|
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static void ss_wakeup(struct list_head *h, int kill)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-05-01 10:15:49 +08:00
|
|
|
struct msg_sender *mss, *t;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-05-01 10:15:49 +08:00
|
|
|
list_for_each_entry_safe(mss, t, h, list) {
|
2006-07-30 18:04:11 +08:00
|
|
|
if (kill)
|
|
|
|
mss->list.next = NULL;
|
2005-04-17 06:20:36 +08:00
|
|
|
wake_up_process(mss->tsk);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static void expunge_all(struct msg_queue *msq, int res)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-05-01 10:15:49 +08:00
|
|
|
struct msg_receiver *msr, *t;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2013-05-01 10:15:49 +08:00
|
|
|
list_for_each_entry_safe(msr, t, &msq->q_receivers, r_list) {
|
2005-04-17 06:20:36 +08:00
|
|
|
msr->r_msg = NULL;
|
|
|
|
wake_up_process(msr->r_tsk);
|
|
|
|
smp_mb();
|
|
|
|
msr->r_msg = ERR_PTR(res);
|
|
|
|
}
|
|
|
|
}
|
2006-07-30 18:04:11 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* freeque() wakes up waiters on the sender and receiver waiting queue,
|
2007-10-19 14:40:53 +08:00
|
|
|
* removes the message queue from message queue ID IDR, and cleans up all the
|
|
|
|
* messages associated with this queue.
|
2005-04-17 06:20:36 +08:00
|
|
|
*
|
2007-10-19 14:40:54 +08:00
|
|
|
* msg_ids.rw_mutex (writer) and the spinlock for this message queue are held
|
|
|
|
* before freeque() is called. msg_ids.rw_mutex remains locked on exit.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2008-02-08 20:18:57 +08:00
|
|
|
static void freeque(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-05-01 10:15:49 +08:00
|
|
|
struct msg_msg *msg, *t;
|
2008-02-08 20:18:57 +08:00
|
|
|
struct msg_queue *msq = container_of(ipcp, struct msg_queue, q_perm);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
expunge_all(msq, -EIDRM);
|
|
|
|
ss_wakeup(&msq->q_senders, 1);
|
2007-10-19 14:40:48 +08:00
|
|
|
msg_rmid(ns, msq);
|
2005-04-17 06:20:36 +08:00
|
|
|
msg_unlock(msq);
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2013-05-01 10:15:49 +08:00
|
|
|
list_for_each_entry_safe(msg, t, &msq->q_messages, m_list) {
|
2007-10-19 14:40:56 +08:00
|
|
|
atomic_dec(&ns->msg_hdrs);
|
2005-04-17 06:20:36 +08:00
|
|
|
free_msg(msg);
|
|
|
|
}
|
2007-10-19 14:40:56 +08:00
|
|
|
atomic_sub(msq->q_cbytes, &ns->msg_bytes);
|
2005-04-17 06:20:36 +08:00
|
|
|
security_msg_queue_free(msq);
|
|
|
|
ipc_rcu_putref(msq);
|
|
|
|
}
|
|
|
|
|
2007-10-19 14:40:53 +08:00
|
|
|
/*
|
2007-10-19 14:40:54 +08:00
|
|
|
* Called with msg_ids.rw_mutex and ipcp locked.
|
2007-10-19 14:40:53 +08:00
|
|
|
*/
|
2007-10-19 14:40:51 +08:00
|
|
|
static inline int msg_security(struct kern_ipc_perm *ipcp, int msgflg)
|
2007-10-19 14:40:49 +08:00
|
|
|
{
|
2007-10-19 14:40:51 +08:00
|
|
|
struct msg_queue *msq = container_of(ipcp, struct msg_queue, q_perm);
|
|
|
|
|
|
|
|
return security_msg_queue_associate(msq, msgflg);
|
2007-10-19 14:40:49 +08:00
|
|
|
}
|
|
|
|
|
2009-01-14 21:14:26 +08:00
|
|
|
SYSCALL_DEFINE2(msgget, key_t, key, int, msgflg)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2006-10-02 17:18:21 +08:00
|
|
|
struct ipc_namespace *ns;
|
2007-10-19 14:40:49 +08:00
|
|
|
struct ipc_ops msg_ops;
|
|
|
|
struct ipc_params msg_params;
|
2006-10-02 17:18:21 +08:00
|
|
|
|
|
|
|
ns = current->nsproxy->ipc_ns;
|
2007-10-19 14:40:48 +08:00
|
|
|
|
2007-10-19 14:40:49 +08:00
|
|
|
msg_ops.getnew = newque;
|
|
|
|
msg_ops.associate = msg_security;
|
|
|
|
msg_ops.more_checks = NULL;
|
|
|
|
|
|
|
|
msg_params.key = key;
|
|
|
|
msg_params.flg = msgflg;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2007-10-19 14:40:49 +08:00
|
|
|
return ipcget(ns, &msg_ids(ns), &msg_ops, &msg_params);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static inline unsigned long
|
|
|
|
copy_msqid_to_user(void __user *buf, struct msqid64_ds *in, int version)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
switch(version) {
|
|
|
|
case IPC_64:
|
2006-07-30 18:04:11 +08:00
|
|
|
return copy_to_user(buf, in, sizeof(*in));
|
2005-04-17 06:20:36 +08:00
|
|
|
case IPC_OLD:
|
2006-07-30 18:04:11 +08:00
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
struct msqid_ds out;
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
memset(&out, 0, sizeof(out));
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
ipc64_perm_to_ipc_perm(&in->msg_perm, &out.msg_perm);
|
|
|
|
|
|
|
|
out.msg_stime = in->msg_stime;
|
|
|
|
out.msg_rtime = in->msg_rtime;
|
|
|
|
out.msg_ctime = in->msg_ctime;
|
|
|
|
|
2010-05-25 05:33:03 +08:00
|
|
|
if (in->msg_cbytes > USHRT_MAX)
|
|
|
|
out.msg_cbytes = USHRT_MAX;
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
|
|
|
out.msg_cbytes = in->msg_cbytes;
|
|
|
|
out.msg_lcbytes = in->msg_cbytes;
|
|
|
|
|
2010-05-25 05:33:03 +08:00
|
|
|
if (in->msg_qnum > USHRT_MAX)
|
|
|
|
out.msg_qnum = USHRT_MAX;
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
|
|
|
out.msg_qnum = in->msg_qnum;
|
|
|
|
|
2010-05-25 05:33:03 +08:00
|
|
|
if (in->msg_qbytes > USHRT_MAX)
|
|
|
|
out.msg_qbytes = USHRT_MAX;
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
|
|
|
out.msg_qbytes = in->msg_qbytes;
|
|
|
|
out.msg_lqbytes = in->msg_qbytes;
|
|
|
|
|
|
|
|
out.msg_lspid = in->msg_lspid;
|
|
|
|
out.msg_lrpid = in->msg_lrpid;
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
return copy_to_user(buf, &out, sizeof(out));
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static inline unsigned long
|
2008-04-29 16:00:50 +08:00
|
|
|
copy_msqid_from_user(struct msqid64_ds *out, void __user *buf, int version)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
switch(version) {
|
|
|
|
case IPC_64:
|
2008-04-29 16:00:50 +08:00
|
|
|
if (copy_from_user(out, buf, sizeof(*out)))
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EFAULT;
|
|
|
|
return 0;
|
|
|
|
case IPC_OLD:
|
2006-07-30 18:04:11 +08:00
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
struct msqid_ds tbuf_old;
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
if (copy_from_user(&tbuf_old, buf, sizeof(tbuf_old)))
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EFAULT;
|
|
|
|
|
2008-04-29 16:00:50 +08:00
|
|
|
out->msg_perm.uid = tbuf_old.msg_perm.uid;
|
|
|
|
out->msg_perm.gid = tbuf_old.msg_perm.gid;
|
|
|
|
out->msg_perm.mode = tbuf_old.msg_perm.mode;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
if (tbuf_old.msg_qbytes == 0)
|
2008-04-29 16:00:50 +08:00
|
|
|
out->msg_qbytes = tbuf_old.msg_lqbytes;
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
2008-04-29 16:00:50 +08:00
|
|
|
out->msg_qbytes = tbuf_old.msg_qbytes;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
return 0;
|
2006-07-30 18:04:11 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-04-29 16:00:48 +08:00
|
|
|
/*
|
|
|
|
* This function handles some msgctl commands which require the rw_mutex
|
|
|
|
* to be held in write mode.
|
|
|
|
* NOTE: no locks must be held, the rw_mutex is taken inside this function.
|
|
|
|
*/
|
|
|
|
static int msgctl_down(struct ipc_namespace *ns, int msqid, int cmd,
|
|
|
|
struct msqid_ds __user *buf, int version)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct kern_ipc_perm *ipcp;
|
2009-10-19 06:54:29 +08:00
|
|
|
struct msqid64_ds uninitialized_var(msqid64);
|
2008-04-29 16:00:48 +08:00
|
|
|
struct msg_queue *msq;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (cmd == IPC_SET) {
|
2008-04-29 16:00:50 +08:00
|
|
|
if (copy_msqid_from_user(&msqid64, buf, version))
|
2008-04-29 16:00:48 +08:00
|
|
|
return -EFAULT;
|
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:12 +08:00
|
|
|
down_write(&msg_ids(ns).rw_mutex);
|
|
|
|
rcu_read_lock();
|
|
|
|
|
2013-07-09 07:01:13 +08:00
|
|
|
ipcp = ipcctl_pre_down_nolock(ns, &msg_ids(ns), msqid, cmd,
|
|
|
|
&msqid64.msg_perm, msqid64.msg_qbytes);
|
2013-07-09 07:01:12 +08:00
|
|
|
if (IS_ERR(ipcp)) {
|
|
|
|
err = PTR_ERR(ipcp);
|
|
|
|
goto out_unlock1;
|
|
|
|
}
|
2008-04-29 16:00:48 +08:00
|
|
|
|
2008-04-29 16:00:54 +08:00
|
|
|
msq = container_of(ipcp, struct msg_queue, q_perm);
|
2008-04-29 16:00:48 +08:00
|
|
|
|
|
|
|
err = security_msg_queue_msgctl(msq, cmd);
|
|
|
|
if (err)
|
2013-07-09 07:01:13 +08:00
|
|
|
goto out_unlock1;
|
2008-04-29 16:00:48 +08:00
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
case IPC_RMID:
|
2013-07-09 07:01:13 +08:00
|
|
|
ipc_lock_object(&msq->q_perm);
|
2013-07-09 07:01:12 +08:00
|
|
|
/* freeque unlocks the ipc object and rcu */
|
2008-04-29 16:00:48 +08:00
|
|
|
freeque(ns, ipcp);
|
|
|
|
goto out_up;
|
|
|
|
case IPC_SET:
|
2008-04-29 16:00:50 +08:00
|
|
|
if (msqid64.msg_qbytes > ns->msg_ctlmnb &&
|
2008-04-29 16:00:48 +08:00
|
|
|
!capable(CAP_SYS_RESOURCE)) {
|
|
|
|
err = -EPERM;
|
2013-07-09 07:01:13 +08:00
|
|
|
goto out_unlock1;
|
2008-04-29 16:00:48 +08:00
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:13 +08:00
|
|
|
ipc_lock_object(&msq->q_perm);
|
2012-02-08 08:54:11 +08:00
|
|
|
err = ipc_update_perm(&msqid64.msg_perm, ipcp);
|
|
|
|
if (err)
|
2013-07-09 07:01:12 +08:00
|
|
|
goto out_unlock0;
|
2012-02-08 08:54:11 +08:00
|
|
|
|
2008-04-29 16:00:50 +08:00
|
|
|
msq->q_qbytes = msqid64.msg_qbytes;
|
2008-04-29 16:00:48 +08:00
|
|
|
|
|
|
|
msq->q_ctime = get_seconds();
|
|
|
|
/* sleeping receivers might be excluded by
|
|
|
|
* stricter permissions.
|
|
|
|
*/
|
|
|
|
expunge_all(msq, -EAGAIN);
|
|
|
|
/* sleeping senders might be able to send
|
|
|
|
* due to a larger queue size.
|
|
|
|
*/
|
|
|
|
ss_wakeup(&msq->q_senders, 0);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
err = -EINVAL;
|
2013-07-09 07:01:13 +08:00
|
|
|
goto out_unlock1;
|
2008-04-29 16:00:48 +08:00
|
|
|
}
|
2013-07-09 07:01:12 +08:00
|
|
|
|
|
|
|
out_unlock0:
|
|
|
|
ipc_unlock_object(&msq->q_perm);
|
|
|
|
out_unlock1:
|
|
|
|
rcu_read_unlock();
|
2008-04-29 16:00:48 +08:00
|
|
|
out_up:
|
|
|
|
up_write(&msg_ids(ns).rw_mutex);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:14 +08:00
|
|
|
static int msgctl_nolock(struct ipc_namespace *ns, int msqid,
|
|
|
|
int cmd, int version, void __user *buf)
|
2008-04-29 16:00:48 +08:00
|
|
|
{
|
2013-07-09 07:01:14 +08:00
|
|
|
int err;
|
2006-07-30 18:04:11 +08:00
|
|
|
struct msg_queue *msq;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
switch (cmd) {
|
2006-07-30 18:04:11 +08:00
|
|
|
case IPC_INFO:
|
|
|
|
case MSG_INFO:
|
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
struct msginfo msginfo;
|
|
|
|
int max_id;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (!buf)
|
|
|
|
return -EFAULT;
|
2013-07-09 07:01:14 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
/*
|
|
|
|
* We must not return kernel stack data.
|
2005-04-17 06:20:36 +08:00
|
|
|
* due to padding, it's not enough
|
|
|
|
* to set all member fields.
|
|
|
|
*/
|
|
|
|
err = security_msg_queue_msgctl(NULL, cmd);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
memset(&msginfo, 0, sizeof(msginfo));
|
2006-10-02 17:18:21 +08:00
|
|
|
msginfo.msgmni = ns->msg_ctlmni;
|
|
|
|
msginfo.msgmax = ns->msg_ctlmax;
|
|
|
|
msginfo.msgmnb = ns->msg_ctlmnb;
|
2005-04-17 06:20:36 +08:00
|
|
|
msginfo.msgssz = MSGSSZ;
|
|
|
|
msginfo.msgseg = MSGSEG;
|
2007-10-19 14:40:54 +08:00
|
|
|
down_read(&msg_ids(ns).rw_mutex);
|
2005-04-17 06:20:36 +08:00
|
|
|
if (cmd == MSG_INFO) {
|
2006-10-02 17:18:21 +08:00
|
|
|
msginfo.msgpool = msg_ids(ns).in_use;
|
2007-10-19 14:40:56 +08:00
|
|
|
msginfo.msgmap = atomic_read(&ns->msg_hdrs);
|
|
|
|
msginfo.msgtql = atomic_read(&ns->msg_bytes);
|
2005-04-17 06:20:36 +08:00
|
|
|
} else {
|
|
|
|
msginfo.msgmap = MSGMAP;
|
|
|
|
msginfo.msgpool = MSGPOOL;
|
|
|
|
msginfo.msgtql = MSGTQL;
|
|
|
|
}
|
2007-10-19 14:40:48 +08:00
|
|
|
max_id = ipc_get_maxid(&msg_ids(ns));
|
2007-10-19 14:40:54 +08:00
|
|
|
up_read(&msg_ids(ns).rw_mutex);
|
2006-07-30 18:04:11 +08:00
|
|
|
if (copy_to_user(buf, &msginfo, sizeof(struct msginfo)))
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EFAULT;
|
2006-07-30 18:04:11 +08:00
|
|
|
return (max_id < 0) ? 0 : max_id;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2013-07-09 07:01:14 +08:00
|
|
|
|
|
|
|
case MSG_STAT:
|
2005-04-17 06:20:36 +08:00
|
|
|
case IPC_STAT:
|
|
|
|
{
|
|
|
|
struct msqid64_ds tbuf;
|
|
|
|
int success_return;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (!buf)
|
|
|
|
return -EFAULT;
|
|
|
|
|
2013-07-09 07:01:16 +08:00
|
|
|
memset(&tbuf, 0, sizeof(tbuf));
|
|
|
|
|
|
|
|
rcu_read_lock();
|
2006-07-30 18:04:11 +08:00
|
|
|
if (cmd == MSG_STAT) {
|
2013-07-09 07:01:16 +08:00
|
|
|
msq = msq_obtain_object(ns, msqid);
|
|
|
|
if (IS_ERR(msq)) {
|
|
|
|
err = PTR_ERR(msq);
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
2007-10-19 14:40:48 +08:00
|
|
|
success_return = msq->q_perm.id;
|
2005-04-17 06:20:36 +08:00
|
|
|
} else {
|
2013-07-09 07:01:16 +08:00
|
|
|
msq = msq_obtain_object_check(ns, msqid);
|
|
|
|
if (IS_ERR(msq)) {
|
|
|
|
err = PTR_ERR(msq);
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
success_return = 0;
|
|
|
|
}
|
2013-07-09 07:01:16 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
err = -EACCES;
|
2011-03-24 07:43:24 +08:00
|
|
|
if (ipcperms(ns, &msq->q_perm, S_IRUGO))
|
2005-04-17 06:20:36 +08:00
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
err = security_msg_queue_msgctl(msq, cmd);
|
|
|
|
if (err)
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
kernel_to_ipc64_perm(&msq->q_perm, &tbuf.msg_perm);
|
|
|
|
tbuf.msg_stime = msq->q_stime;
|
|
|
|
tbuf.msg_rtime = msq->q_rtime;
|
|
|
|
tbuf.msg_ctime = msq->q_ctime;
|
|
|
|
tbuf.msg_cbytes = msq->q_cbytes;
|
|
|
|
tbuf.msg_qnum = msq->q_qnum;
|
|
|
|
tbuf.msg_qbytes = msq->q_qbytes;
|
|
|
|
tbuf.msg_lspid = msq->q_lspid;
|
|
|
|
tbuf.msg_lrpid = msq->q_lrpid;
|
2013-07-09 07:01:16 +08:00
|
|
|
rcu_read_unlock();
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (copy_msqid_to_user(buf, &tbuf, version))
|
|
|
|
return -EFAULT;
|
|
|
|
return success_return;
|
|
|
|
}
|
2013-07-09 07:01:14 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
default:
|
2013-07-09 07:01:14 +08:00
|
|
|
return -EINVAL;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:14 +08:00
|
|
|
return err;
|
2005-04-17 06:20:36 +08:00
|
|
|
out_unlock:
|
2013-07-09 07:01:16 +08:00
|
|
|
rcu_read_unlock();
|
2005-04-17 06:20:36 +08:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:14 +08:00
|
|
|
SYSCALL_DEFINE3(msgctl, int, msqid, int, cmd, struct msqid_ds __user *, buf)
|
|
|
|
{
|
|
|
|
int version;
|
|
|
|
struct ipc_namespace *ns;
|
|
|
|
|
|
|
|
if (msqid < 0 || cmd < 0)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
version = ipc_parse_version(&cmd);
|
|
|
|
ns = current->nsproxy->ipc_ns;
|
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
case IPC_INFO:
|
|
|
|
case MSG_INFO:
|
|
|
|
case MSG_STAT: /* msqid is an index rather than a msg queue id */
|
|
|
|
case IPC_STAT:
|
|
|
|
return msgctl_nolock(ns, msqid, cmd, version, buf);
|
|
|
|
case IPC_SET:
|
|
|
|
case IPC_RMID:
|
|
|
|
return msgctl_down(ns, msqid, cmd, buf, version);
|
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static int testmsg(struct msg_msg *msg, long type, int mode)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
switch(mode)
|
|
|
|
{
|
|
|
|
case SEARCH_ANY:
|
2013-05-01 10:14:54 +08:00
|
|
|
case SEARCH_NUMBER:
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
|
|
|
case SEARCH_LESSEQUAL:
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msg->m_type <=type)
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
|
|
|
break;
|
|
|
|
case SEARCH_EQUAL:
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msg->m_type == type)
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
|
|
|
break;
|
|
|
|
case SEARCH_NOTEQUAL:
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msg->m_type != type)
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static inline int pipelined_send(struct msg_queue *msq, struct msg_msg *msg)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-05-01 10:15:49 +08:00
|
|
|
struct msg_receiver *msr, *t;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2013-05-01 10:15:49 +08:00
|
|
|
list_for_each_entry_safe(msr, t, &msq->q_receivers, r_list) {
|
2006-07-30 18:04:11 +08:00
|
|
|
if (testmsg(msg, msr->r_msgtype, msr->r_mode) &&
|
|
|
|
!security_msg_queue_msgrcv(msq, msg, msr->r_tsk,
|
|
|
|
msr->r_msgtype, msr->r_mode)) {
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
list_del(&msr->r_list);
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msr->r_maxsize < msg->m_ts) {
|
2005-04-17 06:20:36 +08:00
|
|
|
msr->r_msg = NULL;
|
|
|
|
wake_up_process(msr->r_tsk);
|
|
|
|
smp_mb();
|
|
|
|
msr->r_msg = ERR_PTR(-E2BIG);
|
|
|
|
} else {
|
|
|
|
msr->r_msg = NULL;
|
2007-10-19 14:40:14 +08:00
|
|
|
msq->q_lrpid = task_pid_vnr(msr->r_tsk);
|
2005-04-17 06:20:36 +08:00
|
|
|
msq->q_rtime = get_seconds();
|
|
|
|
wake_up_process(msr->r_tsk);
|
|
|
|
smp_mb();
|
|
|
|
msr->r_msg = msg;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-12-07 12:37:48 +08:00
|
|
|
long do_msgsnd(int msqid, long mtype, void __user *mtext,
|
|
|
|
size_t msgsz, int msgflg)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct msg_queue *msq;
|
|
|
|
struct msg_msg *msg;
|
|
|
|
int err;
|
2006-10-02 17:18:21 +08:00
|
|
|
struct ipc_namespace *ns;
|
|
|
|
|
|
|
|
ns = current->nsproxy->ipc_ns;
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2006-10-02 17:18:21 +08:00
|
|
|
if (msgsz > ns->msg_ctlmax || (long) msgsz < 0 || msqid < 0)
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EINVAL;
|
|
|
|
if (mtype < 1)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2006-12-07 12:37:48 +08:00
|
|
|
msg = load_msg(mtext, msgsz);
|
2006-07-30 18:04:11 +08:00
|
|
|
if (IS_ERR(msg))
|
2005-04-17 06:20:36 +08:00
|
|
|
return PTR_ERR(msg);
|
|
|
|
|
|
|
|
msg->m_type = mtype;
|
|
|
|
msg->m_ts = msgsz;
|
|
|
|
|
2013-07-09 07:01:17 +08:00
|
|
|
rcu_read_lock();
|
|
|
|
msq = msq_obtain_object_check(ns, msqid);
|
2007-10-19 14:40:51 +08:00
|
|
|
if (IS_ERR(msq)) {
|
|
|
|
err = PTR_ERR(msq);
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock1;
|
2007-10-19 14:40:51 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
for (;;) {
|
|
|
|
struct msg_sender s;
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
err = -EACCES;
|
2011-03-24 07:43:24 +08:00
|
|
|
if (ipcperms(ns, &msq->q_perm, S_IWUGO))
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock1;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
err = security_msg_queue_msgsnd(msq, msg, msgflg);
|
|
|
|
if (err)
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock1;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msgsz + msq->q_cbytes <= msq->q_qbytes &&
|
2005-04-17 06:20:36 +08:00
|
|
|
1 + msq->q_qnum <= msq->q_qbytes) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* queue full, wait: */
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msgflg & IPC_NOWAIT) {
|
|
|
|
err = -EAGAIN;
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock1;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2013-07-09 07:01:17 +08:00
|
|
|
|
|
|
|
ipc_lock_object(&msq->q_perm);
|
2005-04-17 06:20:36 +08:00
|
|
|
ss_add(msq, &s);
|
2013-05-01 10:15:44 +08:00
|
|
|
|
|
|
|
if (!ipc_rcu_getref(msq)) {
|
|
|
|
err = -EIDRM;
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock0;
|
2013-05-01 10:15:44 +08:00
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:17 +08:00
|
|
|
ipc_unlock_object(&msq->q_perm);
|
|
|
|
rcu_read_unlock();
|
2005-04-17 06:20:36 +08:00
|
|
|
schedule();
|
|
|
|
|
2013-07-09 07:01:17 +08:00
|
|
|
rcu_read_lock();
|
|
|
|
ipc_lock_object(&msq->q_perm);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
ipc_rcu_putref(msq);
|
|
|
|
if (msq->q_perm.deleted) {
|
|
|
|
err = -EIDRM;
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock0;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2013-07-09 07:01:17 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
ss_del(&s);
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (signal_pending(current)) {
|
2006-07-30 18:04:11 +08:00
|
|
|
err = -ERESTARTNOHAND;
|
2013-07-09 07:01:17 +08:00
|
|
|
goto out_unlock0;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2013-07-09 07:01:17 +08:00
|
|
|
|
|
|
|
ipc_unlock_object(&msq->q_perm);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2013-07-09 07:01:17 +08:00
|
|
|
ipc_lock_object(&msq->q_perm);
|
2007-10-19 14:40:14 +08:00
|
|
|
msq->q_lspid = task_tgid_vnr(current);
|
2005-04-17 06:20:36 +08:00
|
|
|
msq->q_stime = get_seconds();
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
if (!pipelined_send(msq, msg)) {
|
2011-03-31 09:57:33 +08:00
|
|
|
/* no one is waiting for this message, enqueue it */
|
2006-07-30 18:04:11 +08:00
|
|
|
list_add_tail(&msg->m_list, &msq->q_messages);
|
2005-04-17 06:20:36 +08:00
|
|
|
msq->q_cbytes += msgsz;
|
|
|
|
msq->q_qnum++;
|
2007-10-19 14:40:56 +08:00
|
|
|
atomic_add(msgsz, &ns->msg_bytes);
|
|
|
|
atomic_inc(&ns->msg_hdrs);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
err = 0;
|
|
|
|
msg = NULL;
|
|
|
|
|
2013-07-09 07:01:17 +08:00
|
|
|
out_unlock0:
|
|
|
|
ipc_unlock_object(&msq->q_perm);
|
|
|
|
out_unlock1:
|
|
|
|
rcu_read_unlock();
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msg != NULL)
|
2005-04-17 06:20:36 +08:00
|
|
|
free_msg(msg);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2009-01-14 21:14:26 +08:00
|
|
|
SYSCALL_DEFINE4(msgsnd, int, msqid, struct msgbuf __user *, msgp, size_t, msgsz,
|
|
|
|
int, msgflg)
|
2006-12-07 12:37:48 +08:00
|
|
|
{
|
|
|
|
long mtype;
|
|
|
|
|
|
|
|
if (get_user(mtype, &msgp->mtype))
|
|
|
|
return -EFAULT;
|
|
|
|
return do_msgsnd(msqid, mtype, msgp->mtext, msgsz, msgflg);
|
|
|
|
}
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
static inline int convert_mode(long *msgtyp, int msgflg)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-05-01 10:14:54 +08:00
|
|
|
if (msgflg & MSG_COPY)
|
|
|
|
return SEARCH_NUMBER;
|
2006-07-30 18:04:11 +08:00
|
|
|
/*
|
2005-04-17 06:20:36 +08:00
|
|
|
* find message of correct type.
|
|
|
|
* msgtyp = 0 => get first.
|
|
|
|
* msgtyp > 0 => get first message of matching type.
|
2006-07-30 18:04:11 +08:00
|
|
|
* msgtyp < 0 => get message with least type must be < abs(msgtype).
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2006-07-30 18:04:11 +08:00
|
|
|
if (*msgtyp == 0)
|
2005-04-17 06:20:36 +08:00
|
|
|
return SEARCH_ANY;
|
2006-07-30 18:04:11 +08:00
|
|
|
if (*msgtyp < 0) {
|
|
|
|
*msgtyp = -*msgtyp;
|
2005-04-17 06:20:36 +08:00
|
|
|
return SEARCH_LESSEQUAL;
|
|
|
|
}
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msgflg & MSG_EXCEPT)
|
2005-04-17 06:20:36 +08:00
|
|
|
return SEARCH_NOTEQUAL;
|
|
|
|
return SEARCH_EQUAL;
|
|
|
|
}
|
|
|
|
|
2013-01-05 07:34:52 +08:00
|
|
|
static long do_msg_fill(void __user *dest, struct msg_msg *msg, size_t bufsz)
|
|
|
|
{
|
|
|
|
struct msgbuf __user *msgp = dest;
|
|
|
|
size_t msgsz;
|
|
|
|
|
|
|
|
if (put_user(msg->m_type, &msgp->mtype))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
msgsz = (bufsz > msg->m_ts) ? msg->m_ts : bufsz;
|
|
|
|
if (store_msg(msgp->mtext, msg, msgsz))
|
|
|
|
return -EFAULT;
|
|
|
|
return msgsz;
|
|
|
|
}
|
|
|
|
|
2013-01-05 07:34:55 +08:00
|
|
|
#ifdef CONFIG_CHECKPOINT_RESTORE
|
2013-01-05 07:35:03 +08:00
|
|
|
/*
|
|
|
|
* This function creates new kernel message structure, large enough to store
|
|
|
|
* bufsz message bytes.
|
|
|
|
*/
|
2013-05-01 10:14:54 +08:00
|
|
|
static inline struct msg_msg *prepare_copy(void __user *buf, size_t bufsz)
|
2013-01-05 07:34:55 +08:00
|
|
|
{
|
|
|
|
struct msg_msg *copy;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Create dummy message to copy real message to.
|
|
|
|
*/
|
|
|
|
copy = load_msg(buf, bufsz);
|
|
|
|
if (!IS_ERR(copy))
|
|
|
|
copy->m_ts = bufsz;
|
|
|
|
return copy;
|
|
|
|
}
|
|
|
|
|
2013-01-05 07:34:58 +08:00
|
|
|
static inline void free_copy(struct msg_msg *copy)
|
2013-01-05 07:34:55 +08:00
|
|
|
{
|
2013-01-05 07:34:58 +08:00
|
|
|
if (copy)
|
2013-01-05 07:34:55 +08:00
|
|
|
free_msg(copy);
|
|
|
|
}
|
|
|
|
#else
|
2013-05-01 10:14:54 +08:00
|
|
|
static inline struct msg_msg *prepare_copy(void __user *buf, size_t bufsz)
|
2013-01-05 07:35:00 +08:00
|
|
|
{
|
|
|
|
return ERR_PTR(-ENOSYS);
|
|
|
|
}
|
|
|
|
|
2013-01-05 07:34:58 +08:00
|
|
|
static inline void free_copy(struct msg_msg *copy)
|
|
|
|
{
|
|
|
|
}
|
2013-01-05 07:34:55 +08:00
|
|
|
#endif
|
|
|
|
|
2013-05-01 10:15:04 +08:00
|
|
|
static struct msg_msg *find_msg(struct msg_queue *msq, long *msgtyp, int mode)
|
|
|
|
{
|
|
|
|
struct msg_msg *msg;
|
|
|
|
long count = 0;
|
|
|
|
|
|
|
|
list_for_each_entry(msg, &msq->q_messages, m_list) {
|
|
|
|
if (testmsg(msg, *msgtyp, mode) &&
|
|
|
|
!security_msg_queue_msgrcv(msq, msg, current,
|
|
|
|
*msgtyp, mode)) {
|
|
|
|
if (mode == SEARCH_LESSEQUAL && msg->m_type != 1) {
|
|
|
|
*msgtyp = msg->m_type - 1;
|
|
|
|
} else if (mode == SEARCH_NUMBER) {
|
|
|
|
if (*msgtyp == count)
|
|
|
|
return msg;
|
|
|
|
} else
|
|
|
|
return msg;
|
|
|
|
count++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return ERR_PTR(-EAGAIN);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2013-01-05 07:34:52 +08:00
|
|
|
long do_msgrcv(int msqid, void __user *buf, size_t bufsz, long msgtyp,
|
|
|
|
int msgflg,
|
|
|
|
long (*msg_handler)(void __user *, struct msg_msg *, size_t))
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct msg_queue *msq;
|
|
|
|
struct msg_msg *msg;
|
|
|
|
int mode;
|
2006-10-02 17:18:21 +08:00
|
|
|
struct ipc_namespace *ns;
|
2013-01-05 07:34:58 +08:00
|
|
|
struct msg_msg *copy = NULL;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-03-09 04:43:27 +08:00
|
|
|
ns = current->nsproxy->ipc_ns;
|
|
|
|
|
2013-01-05 07:34:52 +08:00
|
|
|
if (msqid < 0 || (long) bufsz < 0)
|
2005-04-17 06:20:36 +08:00
|
|
|
return -EINVAL;
|
2013-01-05 07:34:55 +08:00
|
|
|
if (msgflg & MSG_COPY) {
|
2013-05-01 10:14:54 +08:00
|
|
|
copy = prepare_copy(buf, min_t(size_t, bufsz, ns->msg_ctlmax));
|
2013-01-05 07:34:55 +08:00
|
|
|
if (IS_ERR(copy))
|
|
|
|
return PTR_ERR(copy);
|
|
|
|
}
|
2006-07-30 18:04:11 +08:00
|
|
|
mode = convert_mode(&msgtyp, msgflg);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-10-19 14:40:51 +08:00
|
|
|
msq = msg_lock_check(ns, msqid);
|
2013-01-05 07:34:55 +08:00
|
|
|
if (IS_ERR(msq)) {
|
2013-01-05 07:34:58 +08:00
|
|
|
free_copy(copy);
|
2007-10-19 14:40:51 +08:00
|
|
|
return PTR_ERR(msq);
|
2013-01-05 07:34:55 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
for (;;) {
|
|
|
|
struct msg_receiver msr_d;
|
|
|
|
|
|
|
|
msg = ERR_PTR(-EACCES);
|
2011-03-24 07:43:24 +08:00
|
|
|
if (ipcperms(ns, &msq->q_perm, S_IRUGO))
|
2005-04-17 06:20:36 +08:00
|
|
|
goto out_unlock;
|
|
|
|
|
2013-05-01 10:15:04 +08:00
|
|
|
msg = find_msg(msq, &msgtyp, mode);
|
|
|
|
|
2006-07-30 18:04:11 +08:00
|
|
|
if (!IS_ERR(msg)) {
|
|
|
|
/*
|
|
|
|
* Found a suitable message.
|
|
|
|
* Unlink it from the queue.
|
|
|
|
*/
|
2013-01-05 07:34:52 +08:00
|
|
|
if ((bufsz < msg->m_ts) && !(msgflg & MSG_NOERROR)) {
|
2005-04-17 06:20:36 +08:00
|
|
|
msg = ERR_PTR(-E2BIG);
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
2013-01-05 07:35:03 +08:00
|
|
|
/*
|
|
|
|
* If we are copying, then do not unlink message and do
|
|
|
|
* not update queue parameters.
|
|
|
|
*/
|
2013-05-01 10:14:48 +08:00
|
|
|
if (msgflg & MSG_COPY) {
|
|
|
|
msg = copy_msg(msg, copy);
|
2013-01-05 07:34:55 +08:00
|
|
|
goto out_unlock;
|
2013-05-01 10:14:48 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
list_del(&msg->m_list);
|
|
|
|
msq->q_qnum--;
|
|
|
|
msq->q_rtime = get_seconds();
|
2007-10-19 14:40:14 +08:00
|
|
|
msq->q_lrpid = task_tgid_vnr(current);
|
2005-04-17 06:20:36 +08:00
|
|
|
msq->q_cbytes -= msg->m_ts;
|
2007-10-19 14:40:56 +08:00
|
|
|
atomic_sub(msg->m_ts, &ns->msg_bytes);
|
|
|
|
atomic_dec(&ns->msg_hdrs);
|
2006-07-30 18:04:11 +08:00
|
|
|
ss_wakeup(&msq->q_senders, 0);
|
2005-04-17 06:20:36 +08:00
|
|
|
msg_unlock(msq);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
/* No message waiting. Wait for a message */
|
|
|
|
if (msgflg & IPC_NOWAIT) {
|
|
|
|
msg = ERR_PTR(-ENOMSG);
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
2006-07-30 18:04:11 +08:00
|
|
|
list_add_tail(&msr_d.r_list, &msq->q_receivers);
|
2005-04-17 06:20:36 +08:00
|
|
|
msr_d.r_tsk = current;
|
|
|
|
msr_d.r_msgtype = msgtyp;
|
|
|
|
msr_d.r_mode = mode;
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msgflg & MSG_NOERROR)
|
2005-04-17 06:20:36 +08:00
|
|
|
msr_d.r_maxsize = INT_MAX;
|
2006-07-30 18:04:11 +08:00
|
|
|
else
|
2013-01-05 07:34:52 +08:00
|
|
|
msr_d.r_maxsize = bufsz;
|
2005-04-17 06:20:36 +08:00
|
|
|
msr_d.r_msg = ERR_PTR(-EAGAIN);
|
|
|
|
current->state = TASK_INTERRUPTIBLE;
|
|
|
|
msg_unlock(msq);
|
|
|
|
|
|
|
|
schedule();
|
|
|
|
|
|
|
|
/* Lockless receive, part 1:
|
|
|
|
* Disable preemption. We don't hold a reference to the queue
|
|
|
|
* and getting a reference would defeat the idea of a lockless
|
|
|
|
* operation, thus the code relies on rcu to guarantee the
|
2011-03-31 09:57:33 +08:00
|
|
|
* existence of msq:
|
2005-04-17 06:20:36 +08:00
|
|
|
* Prior to destruction, expunge_all(-EIRDM) changes r_msg.
|
|
|
|
* Thus if r_msg is -EAGAIN, then the queue not yet destroyed.
|
|
|
|
* rcu_read_lock() prevents preemption between reading r_msg
|
|
|
|
* and the spin_lock() inside ipc_lock_by_ptr().
|
|
|
|
*/
|
|
|
|
rcu_read_lock();
|
|
|
|
|
|
|
|
/* Lockless receive, part 2:
|
|
|
|
* Wait until pipelined_send or expunge_all are outside of
|
|
|
|
* wake_up_process(). There is a race with exit(), see
|
|
|
|
* ipc/mqueue.c for the details.
|
|
|
|
*/
|
2006-07-30 18:04:11 +08:00
|
|
|
msg = (struct msg_msg*)msr_d.r_msg;
|
2005-04-17 06:20:36 +08:00
|
|
|
while (msg == NULL) {
|
|
|
|
cpu_relax();
|
2006-07-30 18:04:11 +08:00
|
|
|
msg = (struct msg_msg *)msr_d.r_msg;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Lockless receive, part 3:
|
|
|
|
* If there is a message or an error then accept it without
|
|
|
|
* locking.
|
|
|
|
*/
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msg != ERR_PTR(-EAGAIN)) {
|
2005-04-17 06:20:36 +08:00
|
|
|
rcu_read_unlock();
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Lockless receive, part 3:
|
|
|
|
* Acquire the queue spinlock.
|
|
|
|
*/
|
|
|
|
ipc_lock_by_ptr(&msq->q_perm);
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
|
|
|
/* Lockless receive, part 4:
|
|
|
|
* Repeat test after acquiring the spinlock.
|
|
|
|
*/
|
|
|
|
msg = (struct msg_msg*)msr_d.r_msg;
|
2006-07-30 18:04:11 +08:00
|
|
|
if (msg != ERR_PTR(-EAGAIN))
|
2005-04-17 06:20:36 +08:00
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
list_del(&msr_d.r_list);
|
|
|
|
if (signal_pending(current)) {
|
|
|
|
msg = ERR_PTR(-ERESTARTNOHAND);
|
|
|
|
out_unlock:
|
|
|
|
msg_unlock(msq);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2013-01-05 07:34:55 +08:00
|
|
|
if (IS_ERR(msg)) {
|
2013-01-05 07:34:58 +08:00
|
|
|
free_copy(copy);
|
2006-07-30 18:04:11 +08:00
|
|
|
return PTR_ERR(msg);
|
2013-01-05 07:34:55 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-01-05 07:34:52 +08:00
|
|
|
bufsz = msg_handler(buf, msg, bufsz);
|
2005-04-17 06:20:36 +08:00
|
|
|
free_msg(msg);
|
2006-07-30 18:04:11 +08:00
|
|
|
|
2013-01-05 07:34:52 +08:00
|
|
|
return bufsz;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2009-01-14 21:14:26 +08:00
|
|
|
SYSCALL_DEFINE5(msgrcv, int, msqid, struct msgbuf __user *, msgp, size_t, msgsz,
|
|
|
|
long, msgtyp, int, msgflg)
|
2006-12-07 12:37:48 +08:00
|
|
|
{
|
2013-01-05 07:34:52 +08:00
|
|
|
return do_msgrcv(msqid, msgp, msgsz, msgtyp, msgflg, do_msg_fill);
|
2006-12-07 12:37:48 +08:00
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
#ifdef CONFIG_PROC_FS
|
2005-09-07 06:17:10 +08:00
|
|
|
static int sysvipc_msg_proc_show(struct seq_file *s, void *it)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2012-02-08 08:54:11 +08:00
|
|
|
struct user_namespace *user_ns = seq_user_ns(s);
|
2005-09-07 06:17:10 +08:00
|
|
|
struct msg_queue *msq = it;
|
|
|
|
|
|
|
|
return seq_printf(s,
|
2006-07-30 18:04:11 +08:00
|
|
|
"%10d %10d %4o %10lu %10lu %5u %5u %5u %5u %5u %5u %10lu %10lu %10lu\n",
|
|
|
|
msq->q_perm.key,
|
2007-10-19 14:40:48 +08:00
|
|
|
msq->q_perm.id,
|
2006-07-30 18:04:11 +08:00
|
|
|
msq->q_perm.mode,
|
|
|
|
msq->q_cbytes,
|
|
|
|
msq->q_qnum,
|
|
|
|
msq->q_lspid,
|
|
|
|
msq->q_lrpid,
|
2012-02-08 08:54:11 +08:00
|
|
|
from_kuid_munged(user_ns, msq->q_perm.uid),
|
|
|
|
from_kgid_munged(user_ns, msq->q_perm.gid),
|
|
|
|
from_kuid_munged(user_ns, msq->q_perm.cuid),
|
|
|
|
from_kgid_munged(user_ns, msq->q_perm.cgid),
|
2006-07-30 18:04:11 +08:00
|
|
|
msq->q_stime,
|
|
|
|
msq->q_rtime,
|
|
|
|
msq->q_ctime);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
#endif
|