Skip to contentSkip to conversation

ipc: Limit sysctl value to IPCMNI

PATCH In review Takashi Iwai 3 participants 10 messages Jun 8, 2018 – Jun 18, 2018

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipc: Limit sysctl value to IPCMNI

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

---
 ipc/ipc_sysctl.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
ipc/ipc_sysctl.c +4 -1
index 8ad93c29f511..0c8f89f54e61 100644
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
 static struct ctl_table ipc_kern_table[] = {
 	{
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",
Signature · 1 line
2.17.0
#1 lore raw
Quoted 6 lines · On Fri, 8 Jun 2018 15:49:49 +0200 Takashi Iwai <[email protected]> wrote:

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

Waiman Long was working on a (vastly more complicated) patchset to address this.

Quoted 11 lines
ipc/ipc_sysctl.c +1 -0
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
 static struct ctl_table ipc_kern_table[] = {
 	{

ipc/ipc_sysctl.c static struct ctl_table ipc_kern_table[] = { · line 129

Quoted 10 lines
ipc/ipc_sysctl.c +3 -1
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",

What is the back-compatibility situation here?

#2 lore raw
Quoted 7 lines · Andrew Morton
Quoted 4 lines · On Fri, 8 Jun 2018 15:49:49 +0200 Takashi Iwai <[email protected]> wrote:

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

Quoted 5 lines
Quoted 2 lines

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

Waiman Long was working on a (vastly more complicated) patchset to address this.

That's great. Any patch available for testing?

Quoted 8 lines
Quoted 8 lines
ipc/ipc_sysctl.c +1 -0
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
Quoted 15 lines

ipc/ipc_sysctl.c static struct ctl_table ipc_kern_table[] = { · line 129

Quoted 13 lines
ipc/ipc_sysctl.c +3 -1
 static struct ctl_table ipc_kern_table[] = {
 	{
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",

What is the back-compatibility situation here?

It's obviously an error to set such a high value and suppose that it were accepted. So relying on that behavior must be broken in anyway...

thanks,

Takashi

#3 lore raw
Quoted 9 lines · On Sat, 09 Jun 2018 08:48:48 +0200 Takashi Iwai <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 4 lines · On Fri, 8 Jun 2018 15:49:49 +0200 Takashi Iwai <[email protected]> wrote:

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

Waiman Long was working on a (vastly more complicated) patchset to address this.

That's great. Any patch available for testing?

I think
http://lkml.kernel.org/r/[email protected]
is the most recent version.

Quoted 25 lines
Quoted 9 lines
Quoted 9 lines
ipc/ipc_sysctl.c +1 -0
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
Quoted 15 lines

ipc/ipc_sysctl.c static struct ctl_table ipc_kern_table[] = { · line 129

Quoted 13 lines
ipc/ipc_sysctl.c +3 -1
 static struct ctl_table ipc_kern_table[] = {
 	{
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",

What is the back-compatibility situation here?

Quoted 3 lines

It's obviously an error to set such a high value and suppose that it were accepted. So relying on that behavior must be broken in anyway...

Well the present behaviour is to convert higher values downwards, yes?

int ipc_addid(struct ipc_ids *ids, struct kern_ipc_perm *new, int limit)

{
	kuid_t euid;
	kgid_t egid;
	int id, err;

	if (limit > IPCMNI)
		limit = IPCMNI;

So if someone out there is presently setting this to 999999 then their kernel will work just fine. After your proposed change, it will no longer do so - the tuning attempt will fail with -EINVAL.

It really does us no good to say "you shouldn't have been doing that". The fact that they *are* doing it and that it works OK is the kernel developers' fault for not applying suitable checking on day one. I think we're stuck with continuing to accept such input.

#4 lore raw
Quoted 64 lines · Andrew Morton
Quoted 9 lines · On Sat, 09 Jun 2018 08:48:48 +0200 Takashi Iwai <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 4 lines · On Fri, 8 Jun 2018 15:49:49 +0200 Takashi Iwai <[email protected]> wrote:

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

Waiman Long was working on a (vastly more complicated) patchset to address this.

That's great. Any patch available for testing?

I think
http://lkml.kernel.org/r/[email protected]
is the most recent version.

Quoted 25 lines
Quoted 9 lines
Quoted 9 lines
ipc/ipc_sysctl.c +1 -0
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
Quoted 15 lines

ipc/ipc_sysctl.c static struct ctl_table ipc_kern_table[] = { · line 129

Quoted 13 lines
ipc/ipc_sysctl.c +3 -1
 static struct ctl_table ipc_kern_table[] = {
 	{
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",

What is the back-compatibility situation here?

Quoted 3 lines

It's obviously an error to set such a high value and suppose that it were accepted. So relying on that behavior must be broken in anyway...

Well the present behaviour is to convert higher values downwards, yes?

int ipc_addid(struct ipc_ids *ids, struct kern_ipc_perm *new, int limit)

{
	kuid_t euid;
	kgid_t egid;
	int id, err;

	if (limit > IPCMNI)
		limit = IPCMNI;
Quoted 8 lines

So if someone out there is presently setting this to 999999 then their kernel will work just fine. After your proposed change, it will no longer do so - the tuning attempt will fail with -EINVAL.

It really does us no good to say "you shouldn't have been doing that". The fact that they *are* doing it and that it works OK is the kernel developers' fault for not applying suitable checking on day one. I think we're stuck with continuing to accept such input.

Hm, that's one concern, yes.

OTOH, we do secretly ignore the input value, and this isn't what's expected by user, either. Moreover, user-space has no slightest idea which value can be accepted and which not.

Actually I posted it just because of requests from customers who needed to raise the bar, but didn't notice the effect.

Maybe another possible solution would be to add another proc entry to handle this correctly, and make the old one only for compatibility.

thanks,

Takashi

#5 lore raw
Quoted 26 lines · On Tue, 12 Jun 2018 08:36:32 +0200 Takashi Iwai <[email protected]> wrote:
Quoted 11 lines

Well the present behaviour is to convert higher values downwards, yes?

int ipc_addid(struct ipc_ids *ids, struct kern_ipc_perm *new, int limit)

{
	kuid_t euid;
	kgid_t egid;
	int id, err;

	if (limit > IPCMNI)
		limit = IPCMNI;
Quoted 8 lines

So if someone out there is presently setting this to 999999 then their kernel will work just fine. After your proposed change, it will no longer do so - the tuning attempt will fail with -EINVAL.

It really does us no good to say "you shouldn't have been doing that". The fact that they *are* doing it and that it works OK is the kernel developers' fault for not applying suitable checking on day one. I think we're stuck with continuing to accept such input.

Hm, that's one concern, yes.

OTOH, we do secretly ignore the input value, and this isn't what's expected by user, either. Moreover, user-space has no slightest idea which value can be accepted and which not.

Quoted 5 lines

Actually I posted it just because of requests from customers who needed to raise the bar, but didn't notice the effect.

Maybe another possible solution would be to add another proc entry to handle this correctly, and make the old one only for compatibility.

Yes, I guess so. Just leave the old tunable alone. Possibly we could add a printk_once("please switch to the new interface").

#6 lore raw
Quoted 21 lines · Andrew Morton
Quoted 19 lines · On Tue, 12 Jun 2018 08:36:32 +0200 Takashi Iwai <[email protected]> wrote:
Quoted 19 lines

Well the present behaviour is to convert higher values downwards, yes?

int ipc_addid(struct ipc_ids *ids, struct kern_ipc_perm *new, int limit)

{
	kuid_t euid;
	kgid_t egid;
	int id, err;

	if (limit > IPCMNI)
		limit = IPCMNI;

So if someone out there is presently setting this to 999999 then their kernel will work just fine. After your proposed change, it will no longer do so - the tuning attempt will fail with -EINVAL.

It really does us no good to say "you shouldn't have been doing that". The fact that they *are* doing it and that it works OK is the kernel developers' fault for not applying suitable checking on day one. I think we're stuck with continuing to accept such input.

Quoted 8 lines
Quoted 8 lines

Hm, that's one concern, yes.

OTOH, we do secretly ignore the input value, and this isn't what's expected by user, either. Moreover, user-space has no slightest idea which value can be accepted and which not.

Actually I posted it just because of requests from customers who needed to raise the bar, but didn't notice the effect.

My patchset actually contains patch to increase the IPC limit to over 32k.

Quoted 5 lines
Quoted 3 lines

Maybe another possible solution would be to add another proc entry to handle this correctly, and make the old one only for compatibility.

Yes, I guess so. Just leave the old tunable alone. Possibly we could add a printk_once("please switch to the new interface").

A problem with that is the fact that those sysctl parameters have been there for so long that it may be hard to persuade users to use the new sysctl parameters. Earlier version of the patchset had added a clamping minmax function to address that issue. However, it was deemed overkill for adding quite a lot of code with not that much benefit.

Another alternative is to add a new sysctl parameter to turn on compat mode where no error will be reported if it exceeds the limit just like before. So a sysadmin can turn on compat mode if reporting range error is an issue for existing applications.

Cheers,
Longman

#7 lore raw
Quoted 9 lines · On Mon, 18 Jun 2018 13:27:24 +0800 Waiman Long <[email protected]> wrote:
Quoted 5 lines
Quoted 3 lines

Maybe another possible solution would be to add another proc entry to handle this correctly, and make the old one only for compatibility.

Yes, I guess so. Just leave the old tunable alone. Possibly we could add a printk_once("please switch to the new interface").

A problem with that is the fact that those sysctl parameters have been there for so long that it may be hard to persuade users to use the new sysctl parameters.

That's what a once-off warning could help with. Tell them they're using a deprecated interface. It would take a few years, but when people see the old interface is scheduled for removal, they'll switch, over time.

#8 lore raw
In reply to Andrew Morton
Quoted 19 lines · Andrew Morton
Quoted 6 lines · On Fri, 8 Jun 2018 15:49:49 +0200 Takashi Iwai <[email protected]> wrote:

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

Waiman Long was working on a (vastly more complicated) patchset to address this.

Quoted 8 lines
ipc/ipc_sysctl.c +1 -0
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
Quoted 16 lines

ipc/ipc_sysctl.c static struct ctl_table ipc_kern_table[] = { · line 129

Quoted 13 lines
ipc/ipc_sysctl.c +3 -1
 static struct ctl_table ipc_kern_table[] = {
 	{
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",

What is the back-compatibility situation here?

Sorry for the late reply. I am planning to send out an updated patch once the merge window is closed. The latest patch can be found in

https://lkml.org/lkml/2018/5/7/666

Luis has some concern about the use of __read_mostly tag which I am going to remove in the next version.

-Longman

#9 lore raw
Quoted 37 lines · Waiman Long
Quoted 19 lines · Andrew Morton
Quoted 6 lines · On Fri, 8 Jun 2018 15:49:49 +0200 Takashi Iwai <[email protected]> wrote:

Currently shmmni proc entry accepts all entered integer values, but the practical limit is IPCMNI (32768). This confuses user as if a bigger value were accepted but not applied correctly.

This patch changes the proc entry to use *_minmax variant to limit the accepted values accordingly.

Waiman Long was working on a (vastly more complicated) patchset to address this.

Quoted 8 lines
ipc/ipc_sysctl.c +1 -0
@@ -99,6 +99,7 @@ static int proc_ipc_auto_msgmni(struct ctl_table *table, int write,
 static int zero;
 static int one = 1;
 static int int_max = INT_MAX;
+static int ipcmni = IPCMNI;
 
Quoted 16 lines

ipc/ipc_sysctl.c static struct ctl_table ipc_kern_table[] = { · line 129

Quoted 13 lines
ipc/ipc_sysctl.c +3 -1
 static struct ctl_table ipc_kern_table[] = {
 	{
@@ -120,7 +121,9 @@ static struct ctl_table ipc_kern_table[] = {
 		.data		= &init_ipc_ns.shm_ctlmni,
 		.maxlen		= sizeof(init_ipc_ns.shm_ctlmni),
 		.mode		= 0644,
-		.proc_handler	= proc_ipc_dointvec,
+		.proc_handler	= proc_ipc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &ipcmni,
 	},
 	{
 		.procname	= "shm_rmid_forced",

What is the back-compatibility situation here?

Quoted 7 lines

Sorry for the late reply. I am planning to send out an updated patch once the merge window is closed. The latest patch can be found in

https://lkml.org/lkml/2018/5/7/666

Luis has some concern about the use of __read_mostly tag which I am going to remove in the next version.

Thanks, that's as trivial as my patch, unsurprisingly :)

Takashi