Skip to contentSkip to conversation

add umask parameter to procfs

PATCH In review Rene Scharfe 4 participants 4 messages Feb 17, 2005 – Feb 18, 2005

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 add umask parameter to procfs

Add proc.umask kernel parameter. It can be used to restrict permissions on the numerical directories in the root of a proc filesystem, i.e. the directories containing process specific information.

E.g. add proc.umask=077 to your kernel command line and all users except root can only see their own process details (like command line parameters) with ps or top. It can be useful to add a bit of privacy to multi-user servers.

The patch has been inspired by a similar feature in GrSecurity.

It could have also been implemented as a mount option to procfs, but at a higher cost and no apparent benefit -- changes to this umask are not supposed to happen very often. Actually, the previous incarnation of this patch was implemented as a half-assed mount option, but I didn't know then how easy it is to add a kernel parameter.

Comments about usefulness or implementation are very welcome.

Thanks,
Rene

fs/proc/base.c renamed from fs/proc/base.c~ +9 -3
@@ -32,8 +32,14 @@
 #include <linux/mount.h>
 #include <linux/security.h>
 #include <linux/ptrace.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
 #include "internal.h"
 
+static umode_t umask = 0;
+module_param(umask, ushort, 0);
+MODULE_PARM_DESC(umask, "umask for the numerical directories in /proc");
+
 /*
  * For hysterical raisins we keep the same inumbers as in the old procfs.
  * Feel free to change the macro below - just keep the range distinct from
@@ -1369,7 +1375,7 @@ static struct dentry *proc_pident_lookup
 		goto out;
 
 	ei = PROC_I(inode);
-	inode->i_mode = p->mode;
+	inode->i_mode = p->mode & ~(umask & S_IRWXUGO);
 	/*
 	 * Yes, it does not scale. And it should not. Don't add
 	 * new entries into /proc/<tgid>/ without very good reasons.
@@ -1699,7 +1705,7 @@ struct dentry *proc_pid_lookup(struct in
 		put_task_struct(task);
 		goto out;
 	}
-	inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO;
+	inode->i_mode = S_IFDIR | ((S_IRUGO|S_IXUGO) & ~umask);
 	inode->i_op = &proc_tgid_base_inode_operations;
 	inode->i_fop = &proc_tgid_base_operations;
 	inode->i_nlink = 3;
@@ -1754,7 +1760,7 @@ static struct dentry *proc_task_lookup(s
 
 	if (!inode)
 		goto out_drop_task;
-	inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO;
+	inode->i_mode = S_IFDIR | ((S_IRUGO|S_IXUGO) & ~umask);
 	inode->i_op = &proc_tid_base_inode_operations;
 	inode->i_fop = &proc_tid_base_operations;
 	inode->i_nlink = 3;
#1 lore raw
Quoted 10 lines · Rene Scharfe

Add proc.umask kernel parameter. It can be used to restrict permissions on the numerical directories in the root of a proc filesystem, i.e. the directories containing process specific information.

E.g. add proc.umask=077 to your kernel command line and all users except root can only see their own process details (like command line parameters) with ps or top. It can be useful to add a bit of privacy to multi-user servers.

Quoted 7 lines

The patch has been inspired by a similar feature in GrSecurity.

It could have also been implemented as a mount option to procfs, but at a higher cost and no apparent benefit -- changes to this umask are not supposed to happen very often. Actually, the previous incarnation of this patch was implemented as a half-assed mount option, but I didn't know then how easy it is to add a kernel parameter.

The feature seems fairly obscure, although very simple. Is anyone actually likely to use this?

Quoted 2 lines

+static umode_t umask = 0;

a) I think the above should be called proc_umask.

b) You shouldn't initialise it.

c) When adding a kernel parameter you should update Documentation/kernel-parameters.txt

#2 lore raw
Quoted 13 lines · Andrew Morton
Quoted 12 lines · Rene Scharfe

Add proc.umask kernel parameter. It can be used to restrict permissions on the numerical directories in the root of a proc filesystem, i.e. the directories containing process specific information.

E.g. add proc.umask=077 to your kernel command line and all users except root can only see their own process details (like command line parameters) with ps or top. It can be useful to add a bit of privacy to multi-user servers.

The patch has been inspired by a similar feature in GrSecurity.

Quoted 8 lines
Quoted 5 lines

It could have also been implemented as a mount option to procfs, but at a higher cost and no apparent benefit -- changes to this umask are not supposed to happen very often. Actually, the previous incarnation of this patch was implemented as a half-assed mount option, but I didn't know then how easy it is to add a kernel parameter.

The feature seems fairly obscure, although very simple. Is anyone actually likely to use this?

what about parents (and especially the init process) some tools like pstree (or ps in certain cases) depend on their visibility/accessability ...

was this tested except for the trivial case where just plain everything is visible?

what if you want to change it afterwards (when tools did break)?

best,
Herbert

Quoted 13 lines
Quoted 1 line

+static umode_t umask = 0;

a) I think the above should be called proc_umask.

b) You shouldn't initialise it.

c) When adding a kernel parameter you should update Documentation/kernel-parameters.txt
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#3 lore raw
In reply to Andrew Morton
Quoted 2 lines · Andrew Morton

The feature seems fairly obscure, although very simple. Is anyone actually likely to use this?

Yes, it's an obscure obscurity feature. :) There certainly seems to be some interest in it: both GrSecurity and OpenWall contain something similar. They set the mask at compile time, though. But being able to change it at boot time makes it much easier for mainstream distros to include it in their kernels.

Personally I felt the need for the feature when I was a student sharing a server with hundreds others. Sure, it's not very useful on machines offering no shell access at all.

Quoted 8 lines
Quoted 1 line

+static umode_t umask = 0;

a) I think the above should be called proc_umask.

b) You shouldn't initialise it.

c) When adding a kernel parameter you should update Documentation/kernel-parameters.txt

Like below?

Hrm, maybe the option should be named pidumask instead of simply umask?

Documentation/kernel-parameters.txt +7 -4
diff -urp linux-2.6.11-rc4/Documentation/kernel-parameters.txt linux-2.6.11-rc4+/Documentation/kernel-parameters.txt
@@ -1037,16 +1037,19 @@ running once the system is up.
 			[ISAPNP] Exclude memory regions for the autoconfiguration
 			Ranges are in pairs (memory base and size).
 
+	processor.max_cstate=   [HW, ACPI]
+			Limit processor to maximum C-state
+			max_cstate=9 overrides any DMI blacklist limit.
+
+	proc.umask=	[KNL] Restrict permissions of process specific
+			entries in /proc (i.e. the numerical directories).
+
 	profile=	[KNL] Enable kernel profiling via /proc/profile
 			{ schedule | <number> }
 			(param: schedule - profile schedule points}
 			(param: profile step/bucket size as a power of 2 for
 				statistical time based profiling)
 
-	processor.max_cstate=   [HW, ACPI]
-			Limit processor to maximum C-state
-			max_cstate=9 overrides any DMI blacklist limit.
-
 	prompt_ramdisk=	[RAM] List of RAM disks to prompt for floppy disk
 			before loading.
 			See Documentation/ramdisk.txt.
fs/proc/base.c +9 -3
diff -urp linux-2.6.11-rc4/fs/proc/base.c linux-2.6.11-rc4+/fs/proc/base.c
@@ -32,8 +32,14 @@
 #include <linux/mount.h>
 #include <linux/security.h>
 #include <linux/ptrace.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
 #include "internal.h"
 
+static umode_t proc_umask;
+module_param_named(umask, proc_umask, ushort, 0);
+MODULE_PARM_DESC(umask, "umask for the numerical directories in /proc");
+
 /*
  * For hysterical raisins we keep the same inumbers as in the old procfs.
  * Feel free to change the macro below - just keep the range distinct from
@@ -1369,7 +1375,7 @@ static struct dentry *proc_pident_lookup
 		goto out;
 
 	ei = PROC_I(inode);
-	inode->i_mode = p->mode;
+	inode->i_mode = p->mode & ~(proc_umask & S_IALLUGO);
 	/*
 	 * Yes, it does not scale. And it should not. Don't add
 	 * new entries into /proc/<tgid>/ without very good reasons.
@@ -1699,7 +1705,7 @@ struct dentry *proc_pid_lookup(struct in
 		put_task_struct(task);
 		goto out;
 	}
-	inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO;
+	inode->i_mode = S_IFDIR | ((S_IRUGO|S_IXUGO) & ~proc_umask);
 	inode->i_op = &proc_tgid_base_inode_operations;
 	inode->i_fop = &proc_tgid_base_operations;
 	inode->i_nlink = 3;
@@ -1754,7 +1760,7 @@ static struct dentry *proc_task_lookup(s
 
 	if (!inode)
 		goto out_drop_task;
-	inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO;
+	inode->i_mode = S_IFDIR | ((S_IRUGO|S_IXUGO) & ~proc_umask);
 	inode->i_op = &proc_tid_base_inode_operations;
 	inode->i_fop = &proc_tid_base_operations;
 	inode->i_nlink = 3;