Add cgroup support for enabling controllers at boot time

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +3 -0

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct 
 	spin_unlock(&release_list_lock);
 	mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{

Quoted 13 lines · Balbir Singh

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +9 -0

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct 
 	spin_unlock(&release_list_lock);
 	mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{
+	int i;
+	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+		struct cgroup_subsys *ss = subsys[i];
+		if (!strcmp(str, ss->name)) {
+			ss->disabled = 1;
+			break;

kernel/cgroup.c static void cgroup_release_agent(struct · line 3035

Quoted 4 lines

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +4 -0

+		}
+	}
+}
+__setup("cgroup_disable=", cgroup_disable);

This doesn't handle spaces very well, so isn't it possible for the name of a current or future cgroup subsystem to be specified after cgroup_disable= on the command line and have it disabled by accident?

Quoted 7 lines

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +0 -0

@@ -383,6 +383,10 @@ and is between 256 and 4096 characters. 
 	ccw_timeout_log [S390]
 			See Documentation/s390/CommonIO for details.
 

Documentation/kernel-parameters.txt and is between 256 and 4096 characters. · line 389

Quoted 4 lines

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +4 -0

+	cgroup_disable= [KNL] Enable disable a particular controller
+			Format: {name of the controller}
+			See /proc/cgroups for a list of compiled controllers
+

This works on multiple controllers, though, if they follow cgroup_disable=, so the documentation and format should reflect that.

Quoted 16 lines · David Rientjes

Quoted 14 lines · Balbir Singh

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +10 -0

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct 
 	spin_unlock(&release_list_lock);
 	mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{
+	int i;
+	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+		struct cgroup_subsys *ss = subsys[i];
+		if (!strcmp(str, ss->name)) {
+			ss->disabled = 1;
+			break;
+		}

Quoted 8 lines

kernel/cgroup.c static void cgroup_release_agent(struct · line 3035

Quoted 3 lines

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +3 -0

+	}
+}
+__setup("cgroup_disable=", cgroup_disable);

This doesn't handle spaces very well, so isn't it possible for the name of a current or future cgroup subsystem to be specified after cgroup_disable= on the command line and have it disabled by accident?

Quoted 17 lines

Quoted 16 lines · David Rientjes

Quoted 14 lines · Balbir Singh

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +10 -0

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct 
 	spin_unlock(&release_list_lock);
 	mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{
+	int i;
+	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+		struct cgroup_subsys *ss = subsys[i];
+		if (!strcmp(str, ss->name)) {
+			ss->disabled = 1;
+			break;
+		}

Quoted 9 lines

Quoted 8 lines

kernel/cgroup.c static void cgroup_release_agent(struct · line 3035

Quoted 3 lines

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +3 -0

+	}
+}
+__setup("cgroup_disable=", cgroup_disable);

This doesn't handle spaces very well, so isn't it possible for the name of a current or future cgroup subsystem to be specified after cgroup_disable= on the command line and have it disabled by accident?

Hmm, cmdline like

cgroup_disable=cpu,memory, ...

should be written as

cgroup_disable=cpu cgroup_disable=memory ....

Quoted 8 lines

Quoted 8 lines

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +8 -0

+static int __init cgroup_disable(char *str)
+{
+	int i;
+	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+		struct cgroup_subsys *ss = subsys[i];
+		if (!strcmp(str, ss->name)) {
+			ss->disabled = 1;
+			break;

Quoted 9 lines

kernel/cgroup.c static void cgroup_release_agent(struct · line 3035

Quoted 4 lines

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +4 -0

+		}
+	}
+}
+__setup("cgroup_disable=", cgroup_disable);

This doesn't handle spaces very well, so isn't it possible for the name of a current or future cgroup subsystem to be specified after cgroup_disable= on the command line and have it disabled by accident?

How do you distinguish that from the user wanting to disable the controller on purpose? My understanding is that after parsing cgroup_disable=, the rest of the text is passed to cgroup_disable to process further. You'll find that all the __setup() code in the kernel is implemented this way.

Since the command line is logically delimited by spaces, you can accidently disable a subsystem if its name appears in any of your kernel options following your cgroup_disable= option.

Quoted 3 lines

Since the command line is logically delimited by spaces, you can accidently disable a subsystem if its name appears in any of your kernel options following your cgroup_disable= option.

I think that you're confusing this with things like the very early memory init setup parameters, which do operate on the raw commandline.

By the time anything is passed to a __setup() function, it's already been split into separate strings at space boundaries.

Ok, so the cgroup_disable= parameter should be a list of subsystem names delimited by anything other than a space that the user wants disabled. That makes more sense, thanks.

From: Paul Menage <[email protected]>

The effects of cgroup_disable=foo are:

- foo doesn't show up in /proc/cgroups
- foo isn't auto-mounted if you mount all cgroups in a single hierarchy
- foo isn't visible as an individually mountable subsystem

As a result there will only ever be one call to foo->create(), at init time; all processes will stay in this group, and the group will never be mounted on a visible hierarchy. Any additional effects (e.g. not allocating metadata) are up to the foo subsystem.

This doesn't handle early_init subsystems (their "disabled" bit isn't set be, but it could easily be extended to do so if any of the early_init systems wanted it - I think it would just involve some nastier parameter processing since it would occur before the command-line argument parser had been run.

[Balbir added Documentation/kernel-parameters updates]

Signed-off-by Paul Menage <[email protected]>

Signed-off-by Balbir Singh <[email protected]>

---

 Documentation/kernel-parameters.txt |    4 ++++
 include/linux/cgroup.h              |    1 +
 kernel/cgroup.c                     |   27 +++++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 2 deletions(-)

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt

Quoted 37 lines · On Fri, 07 Mar 2008 00:29:52 +0530 Balbir Singh wrote:

From: Paul Menage <[email protected]>

The effects of cgroup_disable=foo are:

- foo doesn't show up in /proc/cgroups
- foo isn't auto-mounted if you mount all cgroups in a single hierarchy
- foo isn't visible as an individually mountable subsystem

As a result there will only ever be one call to foo->create(), at init time; all processes will stay in this group, and the group will never be mounted on a visible hierarchy. Any additional effects (e.g. not allocating metadata) are up to the foo subsystem.

This doesn't handle early_init subsystems (their "disabled" bit isn't set be, but it could easily be extended to do so if any of the early_init systems wanted it - I think it would just involve some nastier parameter processing since it would occur before the command-line argument parser had been run.

[Balbir added Documentation/kernel-parameters updates]

Signed-off-by Paul Menage <[email protected]>

Signed-off-by Balbir Singh <[email protected]>

---

 Documentation/kernel-parameters.txt |    4 ++++
 include/linux/cgroup.h              |    1 +
 kernel/cgroup.c                     |   27 +++++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 2 deletions(-)

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +0 -0

@@ -383,6 +383,10 @@ and is between 256 and 4096 characters. 
 	ccw_timeout_log [S390]
 			See Documentation/s390/CommonIO for details.
 

Documentation/kernel-parameters.txt and is between 256 and 4096 characters. · line 386

Quoted 1 line

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +1 -0

+	cgroup_disable= [KNL] Enable disable a particular controller

So it can enable or disable? or the text has extra text?

Documentation/kernel-parameters.txt and is between 256 and 4096 characters. · line 392

Quoted 6 lines

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +3 -0

+			Format: {name of the controller}
+			See /proc/cgroups for a list of compiled controllers
+
 	checkreqprot	[SELINUX] Set initial checkreqprot flag value.
 			Format: { "0" | "1" }
 			See security/selinux/Kconfig help text.

---
~Randy

Quoted 44 lines · Randy Dunlap

Quoted 42 lines · On Fri, 07 Mar 2008 00:29:52 +0530 Balbir Singh wrote:

From: Paul Menage <[email protected]>

The effects of cgroup_disable=foo are:

- foo doesn't show up in /proc/cgroups
- foo isn't auto-mounted if you mount all cgroups in a single hierarchy
- foo isn't visible as an individually mountable subsystem

As a result there will only ever be one call to foo->create(), at init time; all processes will stay in this group, and the group will never be mounted on a visible hierarchy. Any additional effects (e.g. not allocating metadata) are up to the foo subsystem.

This doesn't handle early_init subsystems (their "disabled" bit isn't set be, but it could easily be extended to do so if any of the early_init
systems wanted it - I think it would just involve some nastier parameter processing since it would occur before the command-line argument parser had been run.

[Balbir added Documentation/kernel-parameters updates]

Signed-off-by Paul Menage <[email protected]>

Signed-off-by Balbir Singh <[email protected]>

---

 Documentation/kernel-parameters.txt |    4 ++++
 include/linux/cgroup.h              |    1 +
 kernel/cgroup.c                     |   27 +++++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 2 deletions(-)

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt
---
linux-2.6.25-rc4/Documentation/kernel-parameters.txt~cgroup_disable

2008-03-06 17:57:32.000000000 +0530
+++ linux-2.6.25-rc4-balbir/Documentation/kernel-parameters.txt   
2008-03-06 18:00:32.000000000 +0530

(quoted diff) +0 -0

@@ -383,6 +383,10 @@ and is between 256 and 4096 characters.     

ccw_timeout_log [S390]
See Documentation/s390/CommonIO for details.

Quoted 4 lines

Quoted 1 line

+ cgroup_disable= [KNL] Enable disable a particular controller

So it can enable or disable? or the text has extra text?

don't think so, should be "Disable a particular controller"

From: Paul Menage <[email protected]>

The effects of cgroup_disable=foo are:

- foo doesn't show up in /proc/cgroups
- foo isn't auto-mounted if you mount all cgroups in a single hierarchy
- foo isn't visible as an individually mountable subsystem

As a result there will only ever be one call to foo->create(), at init time; all processes will stay in this group, and the group will never be mounted on a visible hierarchy. Any additional effects (e.g. not allocating metadata) are up to the foo subsystem.

This doesn't handle early_init subsystems (their "disabled" bit isn't set be, but it could easily be extended to do so if any of the early_init systems wanted it - I think it would just involve some nastier parameter processing since it would occur before the command-line argument parser had been run.

[Balbir added Documentation/kernel-parameters updates]

Signed-off-by Paul Menage <[email protected]>

Signed-off-by Balbir Singh <[email protected]>

---

 Documentation/kernel-parameters.txt |    4 ++++
 include/linux/cgroup.h              |    1 +
 kernel/cgroup.c                     |   27 +++++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 2 deletions(-)

diff -puN include/linux/cgroup.h~cgroup_disable include/linux/cgroup.h

include/linux/cgroup.h renamed from include/linux/cgroup.h~cgroup_disable +1 -0

@@ -256,6 +256,7 @@ struct cgroup_subsys {
 	void (*bind)(struct cgroup_subsys *ss, struct cgroup *root);
 	int subsys_id;
 	int active;
+	int disabled;
 	int early_init;
 #define MAX_CGROUP_TYPE_NAMELEN 32
 	const char *name;

diff -puN kernel/cgroup.c~cgroup_disable kernel/cgroup.c

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +25 -2

@@ -782,7 +782,14 @@ static int parse_cgroupfs_options(char *
 		if (!*token)
 			return -EINVAL;
 		if (!strcmp(token, "all")) {
-			opts->subsys_bits = (1 << CGROUP_SUBSYS_COUNT) - 1;
+			/* Add all non-disabled subsystems */
+			int i;
+			opts->subsys_bits = 0;
+			for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+				struct cgroup_subsys *ss = subsys[i];
+				if (!ss->disabled)
+					opts->subsys_bits |= 1ul << i;
+			}
 		} else if (!strcmp(token, "noprefix")) {
 			set_bit(ROOT_NOPREFIX, &opts->flags);
 		} else if (!strncmp(token, "release_agent=", 14)) {

@@ -800,7 +807,8 @@ static int parse_cgroupfs_options(char *
 			for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
 				ss = subsys[i];
 				if (!strcmp(token, ss->name)) {
-					set_bit(i, &opts->subsys_bits);
+					if (!ss->disabled)
+						set_bit(i, &opts->subsys_bits);
 					break;
 				}
 			}

@@ -2604,6 +2612,8 @@ static int proc_cgroupstats_show(struct 
 	mutex_lock(&cgroup_mutex);
 	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
 		struct cgroup_subsys *ss = subsys[i];
+		if (ss->disabled)
+			continue;
 		seq_printf(m, "%s\t%lu\t%d\n",
 			   ss->name, ss->root->subsys_bits,
 			   ss->root->number_of_cgroups);

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct 
 	spin_unlock(&release_list_lock);
 	mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{
+	int i;
+	for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+		struct cgroup_subsys *ss = subsys[i];
+		if (!strcmp(str, ss->name)) {
+			ss->disabled = 1;
+			break;
+		}
+	}
+}
+__setup("cgroup_disable=", cgroup_disable);

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +0 -0

 	checkreqprot	[SELINUX] Set initial checkreqprot flag value.
 			Format: { "0" | "1" }
 			See security/selinux/Kconfig help text.

_

Quoted 111 lines · Balbir Singh

From: Paul Menage <[email protected]>

The effects of cgroup_disable=foo are:

- foo doesn't show up in /proc/cgroups
- foo isn't auto-mounted if you mount all cgroups in a single hierarchy
- foo isn't visible as an individually mountable subsystem

As a result there will only ever be one call to foo->create(), at init time; all processes will stay in this group, and the group will never be mounted on a visible hierarchy. Any additional effects (e.g. not allocating metadata) are up to the foo subsystem.

This doesn't handle early_init subsystems (their "disabled" bit isn't set be, but it could easily be extended to do so if any of the early_init systems wanted it - I think it would just involve some nastier parameter processing since it would occur before the command-line argument parser had been run.

[Balbir added Documentation/kernel-parameters updates]

Signed-off-by Paul Menage <[email protected]>

Signed-off-by Balbir Singh <[email protected]>

---

 Documentation/kernel-parameters.txt |    4 ++++
 include/linux/cgroup.h              |    1 +
 kernel/cgroup.c                     |   27 +++++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 2 deletions(-)

diff -puN include/linux/cgroup.h~cgroup_disable include/linux/cgroup.h --- linux-2.6.25-rc4/include/linux/cgroup.h~cgroup_disable

2008-03-06 12:19:38.000000000 +0530

+++ linux-2.6.25-rc4-balbir/include/linux/cgroup.h 2008-03-06
12:19:38.000000000 +0530

include/linux/cgroup.h renamed from include/linux/cgroup.h~cgroup_disable +1 -0

@@ -256,6 +256,7 @@ struct cgroup_subsys {
     void (*bind)(struct cgroup_subsys *ss, struct cgroup *root);
     int subsys_id;
     int active;
+    int disabled;
     int early_init;
 #define MAX_CGROUP_TYPE_NAMELEN 32
     const char *name;

diff -puN kernel/cgroup.c~cgroup_disable kernel/cgroup.c
--- linux-2.6.25-rc4/kernel/cgroup.c~cgroup_disable 2008-03-06
12:19:38.000000000 +0530
+++ linux-2.6.25-rc4-balbir/kernel/cgroup.c 2008-03-06
12:19:38.000000000 +0530

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +25 -2

@@ -782,7 +782,14 @@ static int parse_cgroupfs_options(char *
         if (!*token)
             return -EINVAL;
         if (!strcmp(token, "all")) {
-            opts->subsys_bits = (1 << CGROUP_SUBSYS_COUNT) - 1;
+            /* Add all non-disabled subsystems */
+            int i;
+            opts->subsys_bits = 0;
+            for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+                struct cgroup_subsys *ss = subsys[i];
+                if (!ss->disabled)
+                    opts->subsys_bits |= 1ul << i;
+            }
         } else if (!strcmp(token, "noprefix")) {
             set_bit(ROOT_NOPREFIX, &opts->flags);
         } else if (!strncmp(token, "release_agent=", 14)) {

@@ -800,7 +807,8 @@ static int parse_cgroupfs_options(char *
             for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
                 ss = subsys[i];
                 if (!strcmp(token, ss->name)) {
-                    set_bit(i, &opts->subsys_bits);
+                    if (!ss->disabled)
+                        set_bit(i, &opts->subsys_bits);
                     break;
                 }
             }

@@ -2604,6 +2612,8 @@ static int proc_cgroupstats_show(struct     
 mutex_lock(&cgroup_mutex);
     for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
         struct cgroup_subsys *ss = subsys[i];
+        if (ss->disabled)
+            continue;
         seq_printf(m, "%s\t%lu\t%d\n",
                ss->name, ss->root->subsys_bits,
                ss->root->number_of_cgroups);

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct     
 spin_unlock(&release_list_lock);
     mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{
+    int i;
+    for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+        struct cgroup_subsys *ss = subsys[i];
+        if (!strcmp(str, ss->name)) {
+            ss->disabled = 1;
+            break;
+        }
+    }
+}
+__setup("cgroup_disable=", cgroup_disable);

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt
---
linux-2.6.25-rc4/Documentation/kernel-parameters.txt~cgroup_disable

2008-03-06 17:57:32.000000000 +0530
+++ linux-2.6.25-rc4-balbir/Documentation/kernel-parameters.txt   
2008-03-06 18:00:32.000000000 +0530

(quoted diff) +0 -0

@@ -383,6 +383,10 @@ and is between 256 and 4096 characters.     

ccw_timeout_log [S390]
See Documentation/s390/CommonIO for details.

Documentation/kernel-parameters.txt and is between 256 and 4096 characters. · line 389

Quoted 4 lines

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +4 -0

+    cgroup_disable= [KNL] Enable disable a particular controller
+            Format: {name of the controller}
+            See /proc/cgroups for a list of compiled controllers
+

The changelog of this patch:
- foo doesn't show up in /proc/cgroups

So a disabled subsystem won't show up in /proc/cgroups. In a previous mail, I asked whether it will be useful to print out the disable bit in /proc/cgroups, so we can distinguish a subsystem from disaled and not-compiled.

Quoted 119 lines · Li Zefan

Quoted 111 lines · Balbir Singh

From: Paul Menage <[email protected]>

The effects of cgroup_disable=foo are:

- foo doesn't show up in /proc/cgroups
- foo isn't auto-mounted if you mount all cgroups in a single hierarchy
- foo isn't visible as an individually mountable subsystem

As a result there will only ever be one call to foo->create(), at init time; all processes will stay in this group, and the group will never be mounted on a visible hierarchy. Any additional effects (e.g. not allocating metadata) are up to the foo subsystem.

This doesn't handle early_init subsystems (their "disabled" bit isn't set be, but it could easily be extended to do so if any of the early_init systems wanted it - I think it would just involve some nastier parameter processing since it would occur before the command-line argument parser had been run.

[Balbir added Documentation/kernel-parameters updates]

Signed-off-by Paul Menage <[email protected]>

Signed-off-by Balbir Singh <[email protected]>

---

 Documentation/kernel-parameters.txt |    4 ++++
 include/linux/cgroup.h              |    1 +
 kernel/cgroup.c                     |   27 +++++++++++++++++++++++++--
 3 files changed, 30 insertions(+), 2 deletions(-)

diff -puN include/linux/cgroup.h~cgroup_disable include/linux/cgroup.h --- linux-2.6.25-rc4/include/linux/cgroup.h~cgroup_disable

2008-03-06 12:19:38.000000000 +0530

+++ linux-2.6.25-rc4-balbir/include/linux/cgroup.h 2008-03-06
12:19:38.000000000 +0530

include/linux/cgroup.h renamed from include/linux/cgroup.h~cgroup_disable +1 -0

@@ -256,6 +256,7 @@ struct cgroup_subsys {
     void (*bind)(struct cgroup_subsys *ss, struct cgroup *root);
     int subsys_id;
     int active;
+    int disabled;
     int early_init;
 #define MAX_CGROUP_TYPE_NAMELEN 32
     const char *name;

diff -puN kernel/cgroup.c~cgroup_disable kernel/cgroup.c
--- linux-2.6.25-rc4/kernel/cgroup.c~cgroup_disable 2008-03-06
12:19:38.000000000 +0530
+++ linux-2.6.25-rc4-balbir/kernel/cgroup.c 2008-03-06
12:19:38.000000000 +0530

kernel/cgroup.c renamed from kernel/cgroup.c~cgroup_disable +25 -2

@@ -782,7 +782,14 @@ static int parse_cgroupfs_options(char *
         if (!*token)
             return -EINVAL;
         if (!strcmp(token, "all")) {
-            opts->subsys_bits = (1 << CGROUP_SUBSYS_COUNT) - 1;
+            /* Add all non-disabled subsystems */
+            int i;
+            opts->subsys_bits = 0;
+            for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+                struct cgroup_subsys *ss = subsys[i];
+                if (!ss->disabled)
+                    opts->subsys_bits |= 1ul << i;
+            }
         } else if (!strcmp(token, "noprefix")) {
             set_bit(ROOT_NOPREFIX, &opts->flags);
         } else if (!strncmp(token, "release_agent=", 14)) {

@@ -800,7 +807,8 @@ static int parse_cgroupfs_options(char *
             for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
                 ss = subsys[i];
                 if (!strcmp(token, ss->name)) {
-                    set_bit(i, &opts->subsys_bits);
+                    if (!ss->disabled)
+                        set_bit(i, &opts->subsys_bits);
                     break;
                 }
             }

@@ -2604,6 +2612,8 @@ static int proc_cgroupstats_show(struct     
 mutex_lock(&cgroup_mutex);
     for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
         struct cgroup_subsys *ss = subsys[i];
+        if (ss->disabled)
+            continue;
         seq_printf(m, "%s\t%lu\t%d\n",
                ss->name, ss->root->subsys_bits,
                ss->root->number_of_cgroups);

@@ -3010,3 +3020,16 @@ static void cgroup_release_agent(struct     
 spin_unlock(&release_list_lock);
     mutex_unlock(&cgroup_mutex);
 }
+
+static int __init cgroup_disable(char *str)
+{
+    int i;
+    for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
+        struct cgroup_subsys *ss = subsys[i];
+        if (!strcmp(str, ss->name)) {
+            ss->disabled = 1;
+            break;
+        }
+    }
+}
+__setup("cgroup_disable=", cgroup_disable);

diff -puN Documentation/kernel-parameters.txt~cgroup_disable Documentation/kernel-parameters.txt
---
linux-2.6.25-rc4/Documentation/kernel-parameters.txt~cgroup_disable

2008-03-06 17:57:32.000000000 +0530
+++ linux-2.6.25-rc4-balbir/Documentation/kernel-parameters.txt   
2008-03-06 18:00:32.000000000 +0530

(quoted diff) +0 -0

@@ -383,6 +383,10 @@ and is between 256 and 4096 characters.     

ccw_timeout_log [S390]
See Documentation/s390/CommonIO for details.

Documentation/kernel-parameters.txt and is between 256 and 4096 characters. · line 389

Quoted 4 lines

Documentation/kernel-parameters.txt renamed from Documentation/kernel-parameters.txt~cgroup_disable +4 -0

+    cgroup_disable= [KNL] Enable disable a particular controller
+            Format: {name of the controller}
+            See /proc/cgroups for a list of compiled controllers
+

The changelog of this patch:
- foo doesn't show up in /proc/cgroups

Quoted 4 lines

So a disabled subsystem won't show up in /proc/cgroups. In a previous mail, I asked whether it will be useful to print out the disable bit in /proc/cgroups, so we can distinguish a subsystem from disaled and not-compiled.

Hi, Li,

That is a good idea, but can that come in later? We need to get the boot option in, so that users can decide at boot time whether they want the page_container overhead. I'll send out another set of patches to add that feature or work with Paul to see what he thinks about it.