Changing RT priority in kernel 2.6 without CAP_SYS_NICE

Hello

Here is a patch on the permission scheme when changing RT priorities.

Presently, a process without the capability CAP_SYS_NICE can not change its own policy, which is OK.

But it can also not decrease its RT priority (if scheduled with policy SCHED_RR or SCHED_FIFO), which is what this patch changes.

The rationale is the same as for the nice value: a process should be able to require less priority for itself. Increasing the priority is still not allowed.

This is for example useful if you give a multithreaded user process a RT priority, and the process would like to organize its internal threads using priorities also. Then you can give the process the highest priority needed N, and the process starts its threads with lower priorities: N-1, N-2...

The POSIX norm says that the permissions are implementation specific, so I think we can do that.

In a sense, it makes the permissions consistent whatever the policy is: with this patch, process scheduled by SCHED_FIFO, SCHED_RR and SCHED_OTHER can all decrease their priority.

Please tell me what you think!

Regards

Olivier

--- linux-2.6.8-24.11/kernel/sched.c 2005-01-14 16:34:00.000000000
+0100
+++ linux-2.6.8-24.11-sched-patch/kernel/sched.c 2005-04-17
09:27:07.000000000 +0200
@@ -3248,12 +3248,19 @@

                goto out_unlock;

        retval = -EPERM;

- if ((policy == SCHED_FIFO || policy == SCHED_RR) &&
- !capable(CAP_SYS_NICE))
- goto out_unlock;
- if ((current->euid != p->euid) && (current->euid != p->uid) &&
- !capable(CAP_SYS_NICE))
- goto out_unlock;
+ if(!capable(CAP_SYS_NICE)) {
+ /* can't change a policy without cap */
+ if (policy != p->policy)
+ goto out_unlock;
+ /* can't increase priority without cap */
+ if (policy != SCHED_NORMAL &&
+ lp.sched_priority > p->rt_priority)
+ goto out_unlock;
+ /* can't change other processes without cap */
+ if ((current->euid != p->euid) &&
+ (current->euid != p->uid))
+ goto out_unlock;
+ }

        retval = security_task_setscheduler(p, policy, &lp);
        if (retval)

looks fine to me, only minor nits:

- this area of code changed since 2.6.8 so it needed merging.

- whitespace damage: your patch had all tabs as spaces.

- whitespace style: stuff like "if(" should be "if (".

i've reworked and tested the patch (attached below) to apply against the latest scheduler queue in -mm.

Ingo

--
From: Olivier Croquette <[email protected]>

Presently, a process without the capability CAP_SYS_NICE can not change its own policy, which is OK.

But it can also not decrease its RT priority (if scheduled with policy SCHED_RR or SCHED_FIFO), which is what this patch changes.

The rationale is the same as for the nice value: a process should be able to require less priority for itself. Increasing the priority is still not allowed.

This is for example useful if you give a multithreaded user process a RT priority, and the process would like to organize its internal threads using priorities also. Then you can give the process the highest priority needed N, and the process starts its threads with lower
priorities: N-1, N-2...

The POSIX norm says that the permissions are implementation specific, so I think we can do that.

In a sense, it makes the permissions consistent whatever the policy is:
with this patch, process scheduled by SCHED_FIFO, SCHED_RR and SCHED_OTHER can all decrease their priority.

From: Ingo Molnar <[email protected]>

cleaned up and merged to -mm.

@@ -3436,12 +3436,22 @@ recheck:
 	if ((policy == SCHED_NORMAL) != (param->sched_priority == 0))
 		return -EINVAL;
 
-	if ((policy == SCHED_FIFO || policy == SCHED_RR) &&
-	    !capable(CAP_SYS_NICE))
-		return -EPERM;
-	if ((current->euid != p->euid) && (current->euid != p->uid) &&
-	    !capable(CAP_SYS_NICE))
-		return -EPERM;
+	/*
+	 * Allow unprivileged RT tasks to decrease priority:
+	 */
+	if (!capable(CAP_SYS_NICE)) {
+		/* can't change policy */
+		if (policy != p->policy)
+			return -EPERM;
+		/* can't increase priority */
+		if (policy != SCHED_NORMAL &&
+		    param->sched_priority > p->rt_priority)
+			return -EPERM;
+		/* can't change other user's priorities */
+		if ((current->euid != p->euid) &&
+		    (current->euid != p->uid))
+			return -EPERM;
+	}
 
 	retval = security_task_setscheduler(p, policy, param);
 	if (retval)

This patch needed some massaging to copt with the changes in nice-and-rt-prio-rlimits.patch - please check.

I guess we should merge nice-and-rt-prio-rlimits.patch.

@@ -3445,13 +3445,24 @@ recheck:
 	if ((policy == SCHED_NORMAL) != (param->sched_priority == 0))
 		return -EINVAL;
 
-	if ((policy == SCHED_FIFO || policy == SCHED_RR) &&
-	    param->sched_priority > p->signal->rlim[RLIMIT_RTPRIO].rlim_cur &&
-	    !capable(CAP_SYS_NICE))
-		return -EPERM;
-	if ((current->euid != p->euid) && (current->euid != p->uid) &&
-	    !capable(CAP_SYS_NICE))
-		return -EPERM;
+	/*
+	 * Allow unprivileged RT tasks to decrease priority:
+	 */
+	if (!capable(CAP_SYS_NICE)) {
+		/* can't change policy */
+		if (policy != p->policy)
+			return -EPERM;
+		/* can't increase priority */
+		if (policy != SCHED_NORMAL &&
+		    param->sched_priority > p->rt_priority &&
+		    param->sched_priority >
+				p->signal->rlim[RLIMIT_RTPRIO].rlim_cur)
+			return -EPERM;
+		/* can't change other user's priorities */
+		if ((current->euid != p->euid) &&
+		    (current->euid != p->uid))
+			return -EPERM;
+	}
 
 	retval = security_task_setscheduler(p, policy, param);
 	if (retval)

_

Hi!

The patch seems OK to me, I will try to test live in the next days.

Regards

Olivier

the massaging looks ok - and i agree that we should merge the rt-rlimits patch.

Ingo