Skip to contentSkip to conversation

Posix file capabilities in 2.6.24rc2

DISCUSSION Chris Friedhoff 4 participants 13 messages Nov 13, 2007 – Nov 22, 2007

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

Hello,

everything works as expected, but ...

closing X and no capabilities set for xinit does shutdown only the windowmanager and not the X server (Xorg server 1.4) Consolemessage is:
xinit: Operation not permitted (errno 1): Can't kill X server

the xattr capability is removed, when the file is chown'ed.

Chris
--------------------
Chris Friedhoff
[email protected]

#1 lore raw
Quoted 11 lines · Chris Friedhoff

Hello,

everything works as expected, but ...

closing X and no capabilities set for xinit does shutdown only the windowmanager and not the X server (Xorg server 1.4) Consolemessage is:
xinit: Operation not permitted (errno 1): Can't kill X server

the xattr capability is removed, when the file is chown'ed.

Hi Chris,

yes on chown the capability is removed. I'm not quite sure what you're asking? Is your setup depending on being able to chown while keeping file capabilities? Can you give some more details?

thanks,
-serge

#2 lore raw

Hello Serge,

I wanted only to express what I observed.

A "yes it should" confirms its ok.

And yes, I haven't looked into the patches and the name and commentary of file-capabilities-clear-fcaps-on-inode-change.patch explains this already.
I'm preparing to update my page http://www.friedhoff.org/fscaps.html
for 2.6.24, and I also want to explain what one has to take into account or be beware off. If I stumble about this, I think others will also (imho).

I have written a script to change suid binaries and servers, automating the examples I give on the webpage.
In the sequence of commands I was setting fscaps and than chown the binary. Now with the aforementioned patch the fscaps are removed when I chown and the script wasn't working anymore. My point is not my script, it's being surprised and being a bit at a loss. Documenting this helps to clarify things and users to adopt this feature.

The matter with "xinit: Operation not permitted..." happens, when I (unprivileged user) close a from a console started X session. Similar to Andrew Morton'S http://lkml.org/lkml/2006/11/23/15 . The 2.6.24-rc2 kernel has capabilties enabled but /usr/bin/xinit has no capabilities set. It remains the black screen with a cursor, the windowmanager is closed. Is this known? Is this a problem? Does anyone else observes this?
As far as I understand, I dont have to grant / to use capabilities even when the kernel has capabilities enabled.

Chris

Quoted 13 lines · On Tue, 13 Nov 2007 17:53:18 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 11 lines · Chris Friedhoff

Hello,

everything works as expected, but ...

closing X and no capabilities set for xinit does shutdown only the windowmanager and not the X server (Xorg server 1.4) Consolemessage is:
xinit: Operation not permitted (errno 1): Can't kill X server

the xattr capability is removed, when the file is chown'ed.

Quoted 8 lines

Hi Chris,

yes on chown the capability is removed. I'm not quite sure what you're asking? Is your setup depending on being able to chown while keeping file capabilities? Can you give some more details?

thanks,
-serge

--------------------
Chris Friedhoff
[email protected]

#3 lore raw
Quoted 14 lines · Chris Friedhoff

Hello Serge,

I wanted only to express what I observed.

A "yes it should" confirms its ok.

And yes, I haven't looked into the patches and the name and commentary of file-capabilities-clear-fcaps-on-inode-change.patch explains this already.
I'm preparing to update my page http://www.friedhoff.org/fscaps.html
for 2.6.24, and I also want to explain what one has to take into account or be beware off. If I stumble about this, I think others will also (imho).

Quoted 2 lines

I have written a script to change suid binaries and servers, automating the examples I give on the webpage.

Cool, sounds very useful.

Quoted 7 lines

In the sequence of commands I was setting fscaps and than chown the binary. Now with the aforementioned patch the fscaps are removed when I chown and the script wasn't working anymore. My point is not my script, it's being surprised and being a bit at a loss. Documenting this helps to clarify things and users to adopt this feature.

Quoted 7 lines

The matter with "xinit: Operation not permitted..." happens, when I (unprivileged user) close a from a console started X session. Similar to Andrew Morton'S http://lkml.org/lkml/2006/11/23/15 . The 2.6.24-rc2 kernel has capabilties enabled but /usr/bin/xinit has no capabilities set. It remains the black screen with a cursor, the windowmanager is closed. Is this known? Is this a problem? Does anyone else observes this?

I'm setting up a vm to play with this. Will look into it.

Oh, looking at a few branches, I see that the patch for bug# 9247 (on bugzilla.kernel.org) isn't in 2.6.24-rc2 yet. Can you check whether the following patch fixes it?

thanks,
-serge

Quoted 1 line

From 347faf5852644b91632813885784104f4cdb640a Mon Sep 17 00:00:00 2001

From: Serge E. Hallyn <[email protected]>
Date: Wed, 14 Nov 2007 13:00:52 -0500
Subject: [PATCH 1/1] file capabilities: allow sigcont within session (v2.6.24-rc2)

Allow sigcont to be sent to a process with greater capabilities if it is in the same session. Otherwise, a shell from which I've started a root shell and done 'suspend' can't be restarted by the parent shell.

(this patch against v2.6.24-rc2)

---
 security/commoncap.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
security/commoncap.c +2 -0
index bf67871..c9f6867 100644
@@ -534,6 +534,8 @@ int cap_task_kill(struct task_struct *p, struct siginfo *info,
 		 * Used only by usb drivers?
 		 */
 		return 0;
+	if (sig == SIGCONT && (task_session_nr(current)==task_session_nr(p)))
+		return 0;
 	if (cap_issubset(p->cap_permitted, current->cap_permitted))
 		return 0;
 	if (capable(CAP_KILL))
Signature · 1 line
1.5.1.1.GIT
#4 lore raw

No, the patch doesn't fix the problem.
I still have the black screen with the cursor when I close the xsession, only the windowmanager is closed. consolemessage:
xinit: Operation not permitted (errno 1): Can't kill X server kernel has capabilities, xinit has no caps granted.

Chris

Quoted 41 lines

I'm setting up a vm to play with this. Will look into it.

Oh, looking at a few branches, I see that the patch for bug# 9247 (on bugzilla.kernel.org) isn't in 2.6.24-rc2 yet. Can you check whether the following patch fixes it?

thanks,
-serge

Quoted 1 line

From 347faf5852644b91632813885784104f4cdb640a Mon Sep 17 00:00:00 2001

From: Serge E. Hallyn <[email protected]>
Date: Wed, 14 Nov 2007 13:00:52 -0500
Subject: [PATCH 1/1] file capabilities: allow sigcont within session (v2.6.24-rc2)

Allow sigcont to be sent to a process with greater capabilities if it is in the same session. Otherwise, a shell from which I've started a root shell and done 'suspend' can't be restarted by the parent shell.

(this patch against v2.6.24-rc2)

---
 security/commoncap.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
security/commoncap.c +2 -0
index bf67871..c9f6867 100644
@@ -534,6 +534,8 @@ int cap_task_kill(struct task_struct *p, struct siginfo *info,
 		 * Used only by usb drivers?
 		 */
 		return 0;
+	if (sig == SIGCONT && (task_session_nr(current)==task_session_nr(p)))
+		return 0;
 	if (cap_issubset(p->cap_permitted, current->cap_permitted))
 		return 0;
 	if (capable(CAP_KILL))
Signature · 1 line
1.5.1.1.GIT

--------------------
Chris Friedhoff
[email protected]

#5 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Chris

Quoted 48 lines · On Thu, 15 Nov 2007 23:02:27 +0100 Chris Friedhoff <[email protected]> wrote:

No, the patch doesn't fix the problem.
I still have the black screen with the cursor when I close the xsession, only the windowmanager is closed. consolemessage:
xinit: Operation not permitted (errno 1): Can't kill X server kernel has capabilities, xinit has no caps granted.

Chris

Quoted 38 lines

I'm setting up a vm to play with this. Will look into it.

Oh, looking at a few branches, I see that the patch for bug# 9247 (on bugzilla.kernel.org) isn't in 2.6.24-rc2 yet. Can you check whether the following patch fixes it?

thanks,
-serge

Quoted 1 line

From 347faf5852644b91632813885784104f4cdb640a Mon Sep 17 00:00:00 2001

From: Serge E. Hallyn <[email protected]>
Date: Wed, 14 Nov 2007 13:00:52 -0500
Subject: [PATCH 1/1] file capabilities: allow sigcont within session (v2.6.24-rc2)

Allow sigcont to be sent to a process with greater capabilities if it is in the same session. Otherwise, a shell from which I've started a root shell and done 'suspend' can't be restarted by the parent shell.

(this patch against v2.6.24-rc2)

---
 security/commoncap.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
security/commoncap.c +2 -0
index bf67871..c9f6867 100644
@@ -534,6 +534,8 @@ int cap_task_kill(struct task_struct *p, struct siginfo *info,
 		 * Used only by usb drivers?
 		 */
 		return 0;
+	if (sig == SIGCONT && (task_session_nr(current)==task_session_nr(p)))
+		return 0;
 	if (cap_issubset(p->cap_permitted, current->cap_permitted))
 		return 0;
Quoted 8 lines
Quoted 3 lines
security/commoncap.c +0 -0
 	if (capable(CAP_KILL))
Signature · 1 line
1.5.1.1.GIT

--------------------
Chris Friedhoff
[email protected]

--------------------
Chris Friedhoff
[email protected]

#6 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Quoted 3 lines · Chris Friedhoff

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Ok, so here is the flow.

First off, using runlevel 5 on FC7, using 'log out' correctly brings you back to a new login prompt. Your problem is starting in runlevel 3, and typing 'xinit .xinitrc'; when you exit your wm, xinit is not allowed to kill X so you don't get back to your console.

First comment is, as you point out on your homepage, you could
setfcaps -c cap_kill+p -e /usr/bin/xinit
Then xinit is allowed to kill X. Yes xinit forks and execs a user-writable script, but of course upon the exec to start the script cap_kill is lost, so the user can't abuse this.

Since you pointed this out on your homepage, I have to assume you've decided you don't want to give cap_kill to xinit?

My other question is - do we want to maintain this signal restriction? So long as a privileged process isn't dumpable, is it any more dangerous for user hallyn to kill capability-raised process owned by hallyn than it is to kill a setuid process started by hallyn? If we decide no, then maybe we should remove cap_task_kill() as well as the cap_task_setnice(), cap_task_setioprio(), cap_task_setscheduler()?

Or maybe i've just forgotten a compelling scenario...

thanks,
-serge

#7 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Quoted 11 lines · On Mon, 19 Nov 2007 17:16:44 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 3 lines · Chris Friedhoff

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Ok, so here is the flow.

First off, using runlevel 5 on FC7, using 'log out' correctly brings you back to a new login prompt. Your problem is starting in runlevel 3, and typing 'xinit .xinitrc'; when you exit your wm, xinit is not allowed to kill X so you don't get back to your console.

Yes, I'm booting in a runlevel without a session manager and starting my X session with xinit.
(slackware: console->runlevel 3; sessionmanager->runlevel 4 )

Quoted 9 lines

First comment is, as you point out on your homepage, you could
setfcaps -c cap_kill+p -e /usr/bin/xinit
Then xinit is allowed to kill X. Yes xinit forks and execs a user-writable script, but of course upon the exec to start the script cap_kill is lost, so the user can't abuse this.

Since you pointed this out on your homepage, I have to assume you've decided you don't want to give cap_kill to xinit?

No, since I'm using capabilities and I'm very happy with it, I grant cap_kill to xinit. For myself the problem is solved ...

Quoted 8 lines

My other question is - do we want to maintain this signal restriction? So long as a privileged process isn't dumpable, is it any more dangerous for user hallyn to kill capability-raised process owned by hallyn than it is to kill a setuid process started by hallyn? If we decide no, then maybe we should remove cap_task_kill() as well as the cap_task_setnice(), cap_task_setioprio(), cap_task_setscheduler()?

Quoted 4 lines

Or maybe i've just forgotten a compelling scenario...

thanks,
-serge

... but if some user decides to configure capabilities into the 2.6.24 kernel or just uses such a kernel and
1) is not granting cap_kill to xinit, and
2) starts X by issuing xinit on the console
3) ends after some time his X session, to come back to the console

he will see a different behavior compared to 2.6.23 exiting his X session and (I think) believes to have a bug in the X package.

Andrew Morton describes the problem here, too:
http://lkml.org/lkml/2006/11/23/15
http://lkml.org/lkml/2006/11/23/19

Am I wrong in the assumption, but should one not accept an unchanged behavior with or without capabilities in the kernel regarding the behavior of applications, when he is not actually using (by not setting the xattr capability) capabilities with this application?

If I'm wrong, maybe a warning or hint should be given that one has to grant cap_kill to xinit to come back to the console if the X session was started by xinit.

Chris

--------------------
Chris Friedhoff
[email protected]

#8 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Quoted 59 lines · Chris Friedhoff
Quoted 11 lines · On Mon, 19 Nov 2007 17:16:44 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 3 lines · Chris Friedhoff

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Ok, so here is the flow.

First off, using runlevel 5 on FC7, using 'log out' correctly brings you back to a new login prompt. Your problem is starting in runlevel 3, and typing 'xinit .xinitrc'; when you exit your wm, xinit is not allowed to kill X so you don't get back to your console.

Yes, I'm booting in a runlevel without a session manager and starting my X session with xinit.
(slackware: console->runlevel 3; sessionmanager->runlevel 4 )

Quoted 9 lines

First comment is, as you point out on your homepage, you could
setfcaps -c cap_kill+p -e /usr/bin/xinit
Then xinit is allowed to kill X. Yes xinit forks and execs a user-writable script, but of course upon the exec to start the script cap_kill is lost, so the user can't abuse this.

Since you pointed this out on your homepage, I have to assume you've decided you don't want to give cap_kill to xinit?

No, since I'm using capabilities and I'm very happy with it, I grant cap_kill to xinit. For myself the problem is solved ...

Quoted 8 lines

My other question is - do we want to maintain this signal restriction? So long as a privileged process isn't dumpable, is it any more dangerous for user hallyn to kill capability-raised process owned by hallyn than it is to kill a setuid process started by hallyn? If we decide no, then maybe we should remove cap_task_kill() as well as the cap_task_setnice(), cap_task_setioprio(), cap_task_setscheduler()?

Quoted 4 lines

Or maybe i've just forgotten a compelling scenario...

thanks,
-serge

... but if some user decides to configure capabilities into the 2.6.24 kernel or just uses such a kernel and
1) is not granting cap_kill to xinit, and
2) starts X by issuing xinit on the console
3) ends after some time his X session, to come back to the console

he will see a different behavior compared to 2.6.23 exiting his X session and (I think) believes to have a bug in the X package.

Andrew Morton describes the problem here, too:
http://lkml.org/lkml/2006/11/23/15
http://lkml.org/lkml/2006/11/23/19

Quoted 8 lines

Am I wrong in the assumption, but should one not accept an unchanged behavior with or without capabilities in the kernel regarding the behavior of applications, when he is not actually using (by not setting the xattr capability) capabilities with this application?

If I'm wrong, maybe a warning or hint should be given that one has to grant cap_kill to xinit to come back to the console if the X session was started by xinit.

Thanks - yes, I see (I tend to get lost in my testruns). So we're back to trying to do the fix I was trying to do along with the SIGCONT fix a few weeks ago.

The problem is that when you run a setuid binary, its pP and pE are fully raised. The following patch fixes it for me. Chris, does it fix your problem? Andrew, am I again confusing myself and doing something unsafe?

thanks,
-serge

Quoted 1 line

From d0b931776c0c424e583bf736d6a2498be4eccb98 Mon Sep 17 00:00:00 2001

From: Serge E. Hallyn <[email protected]>
Date: Tue, 20 Nov 2007 08:47:35 +0000
Subject: [PATCH 1/1] file capabilities: don't prevent signaling setuid root programs.

When an unprivileged user runs a setuid root program in !SECURE_NOROOT mode, fP, fI, and fE are set full on, so pP' and pE' are full on. Then cap_task_kill() prevents the user from signaling the setuid root task. This is a change in behavior compared to when !CONFIG_SECURITY_FILE_CAPABILITIES.

This patch introduces a special check into cap_task_kill() just to check whether a non-root user is signaling a setuid root program started by the same user. If so, then signal is allowed.

This still leaves open the question of whether we want to go back to allowing users to signal binaries owned by them which had file capabilities set.

---
 security/commoncap.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)
security/commoncap.c +3 -0
index 302e8d0..d20d0a6 100644
@@ -543,6 +543,9 @@ int cap_task_kill(struct task_struct *p, struct siginfo *info,
 	if (capable(CAP_KILL))
 		return 0;
 
+	if (p->euid==0 && p->uid==current->uid)
+		return 0;
+
 	return -EPERM;
 }
 #else
Signature · 1 line
1.5.2.5
#9 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Quoted 69 lines · On Tue, 20 Nov 2007 08:51:06 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 59 lines · Chris Friedhoff
Quoted 11 lines · On Mon, 19 Nov 2007 17:16:44 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 3 lines · Chris Friedhoff

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Ok, so here is the flow.

First off, using runlevel 5 on FC7, using 'log out' correctly brings you back to a new login prompt. Your problem is starting in runlevel 3, and typing 'xinit .xinitrc'; when you exit your wm, xinit is not allowed to kill X so you don't get back to your console.

Yes, I'm booting in a runlevel without a session manager and starting my X session with xinit.
(slackware: console->runlevel 3; sessionmanager->runlevel 4 )

Quoted 9 lines

First comment is, as you point out on your homepage, you could
setfcaps -c cap_kill+p -e /usr/bin/xinit
Then xinit is allowed to kill X. Yes xinit forks and execs a user-writable script, but of course upon the exec to start the script cap_kill is lost, so the user can't abuse this.

Since you pointed this out on your homepage, I have to assume you've decided you don't want to give cap_kill to xinit?

No, since I'm using capabilities and I'm very happy with it, I grant cap_kill to xinit. For myself the problem is solved ...

Quoted 8 lines

My other question is - do we want to maintain this signal restriction? So long as a privileged process isn't dumpable, is it any more dangerous for user hallyn to kill capability-raised process owned by hallyn than it is to kill a setuid process started by hallyn? If we decide no, then maybe we should remove cap_task_kill() as well as the cap_task_setnice(), cap_task_setioprio(), cap_task_setscheduler()?

Quoted 4 lines

Or maybe i've just forgotten a compelling scenario...

thanks,
-serge

... but if some user decides to configure capabilities into the 2.6.24 kernel or just uses such a kernel and
1) is not granting cap_kill to xinit, and
2) starts X by issuing xinit on the console
3) ends after some time his X session, to come back to the console

he will see a different behavior compared to 2.6.23 exiting his X session and (I think) believes to have a bug in the X package.

Andrew Morton describes the problem here, too:
http://lkml.org/lkml/2006/11/23/15
http://lkml.org/lkml/2006/11/23/19

Quoted 8 lines

Am I wrong in the assumption, but should one not accept an unchanged behavior with or without capabilities in the kernel regarding the behavior of applications, when he is not actually using (by not setting the xattr capability) capabilities with this application?

If I'm wrong, maybe a warning or hint should be given that one has to grant cap_kill to xinit to come back to the console if the X session was started by xinit.

Quoted 7 lines

Thanks - yes, I see (I tend to get lost in my testruns). So we're back to trying to do the fix I was trying to do along with the SIGCONT fix a few weeks ago.

The problem is that when you run a setuid binary, its pP and pE are fully raised. The following patch fixes it for me. Chris, does it fix your problem?

Yes, this patch fixes it for me, too.

Thanks,

Chris

Quoted 46 lines

Andrew, am I again confusing myself and doing something
unsafe?

thanks,
-serge

Quoted 1 line

From d0b931776c0c424e583bf736d6a2498be4eccb98 Mon Sep 17 00:00:00 2001

From: Serge E. Hallyn <[email protected]>
Date: Tue, 20 Nov 2007 08:47:35 +0000
Subject: [PATCH 1/1] file capabilities: don't prevent signaling setuid root programs.

When an unprivileged user runs a setuid root program in !SECURE_NOROOT mode, fP, fI, and fE are set full on, so pP' and pE' are full on. Then cap_task_kill() prevents the user from signaling the setuid root task. This is a change in behavior compared to when !CONFIG_SECURITY_FILE_CAPABILITIES.

This patch introduces a special check into cap_task_kill() just to check whether a non-root user is signaling a setuid root program started by the same user. If so, then signal is allowed.

This still leaves open the question of whether we want to go back to allowing users to signal binaries owned by them which had file capabilities set.

---
 security/commoncap.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)
security/commoncap.c +3 -0
index 302e8d0..d20d0a6 100644
@@ -543,6 +543,9 @@ int cap_task_kill(struct task_struct *p, struct siginfo *info,
 	if (capable(CAP_KILL))
 		return 0;
 
+	if (p->euid==0 && p->uid==current->uid)
+		return 0;
+
 	return -EPERM;
 }
 #else
Signature · 1 line
1.5.2.5

--------------------
Chris Friedhoff
[email protected]

#10 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Quoted 80 lines · Chris Friedhoff
Quoted 69 lines · On Tue, 20 Nov 2007 08:51:06 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 59 lines · Chris Friedhoff
Quoted 11 lines · On Mon, 19 Nov 2007 17:16:44 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 3 lines · Chris Friedhoff

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Ok, so here is the flow.

First off, using runlevel 5 on FC7, using 'log out' correctly brings you back to a new login prompt. Your problem is starting in runlevel 3, and typing 'xinit .xinitrc'; when you exit your wm, xinit is not allowed to kill X so you don't get back to your console.

Yes, I'm booting in a runlevel without a session manager and starting my X session with xinit.
(slackware: console->runlevel 3; sessionmanager->runlevel 4 )

Quoted 9 lines

First comment is, as you point out on your homepage, you could
setfcaps -c cap_kill+p -e /usr/bin/xinit
Then xinit is allowed to kill X. Yes xinit forks and execs a user-writable script, but of course upon the exec to start the script cap_kill is lost, so the user can't abuse this.

Since you pointed this out on your homepage, I have to assume you've decided you don't want to give cap_kill to xinit?

No, since I'm using capabilities and I'm very happy with it, I grant cap_kill to xinit. For myself the problem is solved ...

Quoted 8 lines

My other question is - do we want to maintain this signal restriction? So long as a privileged process isn't dumpable, is it any more dangerous for user hallyn to kill capability-raised process owned by hallyn than it is to kill a setuid process started by hallyn? If we decide no, then maybe we should remove cap_task_kill() as well as the cap_task_setnice(), cap_task_setioprio(), cap_task_setscheduler()?

Quoted 4 lines

Or maybe i've just forgotten a compelling scenario...

thanks,
-serge

... but if some user decides to configure capabilities into the 2.6.24 kernel or just uses such a kernel and
1) is not granting cap_kill to xinit, and
2) starts X by issuing xinit on the console
3) ends after some time his X session, to come back to the console

he will see a different behavior compared to 2.6.23 exiting his X session and (I think) believes to have a bug in the X package.

Andrew Morton describes the problem here, too:
http://lkml.org/lkml/2006/11/23/15
http://lkml.org/lkml/2006/11/23/19

Quoted 8 lines

Am I wrong in the assumption, but should one not accept an unchanged behavior with or without capabilities in the kernel regarding the behavior of applications, when he is not actually using (by not setting the xattr capability) capabilities with this application?

If I'm wrong, maybe a warning or hint should be given that one has to grant cap_kill to xinit to come back to the console if the X session was started by xinit.

Quoted 7 lines

Thanks - yes, I see (I tend to get lost in my testruns). So we're back to trying to do the fix I was trying to do along with the SIGCONT fix a few weeks ago.

The problem is that when you run a setuid binary, its pP and pE are fully raised. The following patch fixes it for me. Chris, does it fix your problem?

Quoted 5 lines

Yes, this patch fixes it for me, too.

Thanks,

Chris

Thanks for the valuable testing and bugreports, Chris.

If I don't hear any counter-arguments I'll post the patch in its own thread tomorrow.

-serge

#11 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

Quoted 81 lines · On Tue, 20 Nov 2007 16:51:21 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 80 lines · Chris Friedhoff
Quoted 76 lines · On Tue, 20 Nov 2007 08:51:06 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 59 lines · Chris Friedhoff
Quoted 11 lines · On Mon, 19 Nov 2007 17:16:44 -0600 "Serge E. Hallyn" <[email protected]> wrote:
Quoted 3 lines · Chris Friedhoff

Hello Serge,

just to let you know: with 2.6.24-rc3 I have the same problem.

Ok, so here is the flow.

First off, using runlevel 5 on FC7, using 'log out' correctly brings you back to a new login prompt. Your problem is starting in runlevel 3, and typing 'xinit .xinitrc'; when you exit your wm, xinit is not allowed to kill X so you don't get back to your console.

Yes, I'm booting in a runlevel without a session manager and starting my X session with xinit.
(slackware: console->runlevel 3; sessionmanager->runlevel 4 )

Quoted 9 lines

First comment is, as you point out on your homepage, you could
setfcaps -c cap_kill+p -e /usr/bin/xinit
Then xinit is allowed to kill X. Yes xinit forks and execs a user-writable script, but of course upon the exec to start the script cap_kill is lost, so the user can't abuse this.

Since you pointed this out on your homepage, I have to assume you've decided you don't want to give cap_kill to xinit?

No, since I'm using capabilities and I'm very happy with it, I grant cap_kill to xinit. For myself the problem is solved ...

Quoted 8 lines

My other question is - do we want to maintain this signal restriction? So long as a privileged process isn't dumpable, is it any more dangerous for user hallyn to kill capability-raised process owned by hallyn than it is to kill a setuid process started by hallyn? If we decide no, then maybe we should remove cap_task_kill() as well as the cap_task_setnice(), cap_task_setioprio(), cap_task_setscheduler()?

Quoted 4 lines

Or maybe i've just forgotten a compelling scenario...

thanks,
-serge

... but if some user decides to configure capabilities into the 2.6.24 kernel or just uses such a kernel and
1) is not granting cap_kill to xinit, and
2) starts X by issuing xinit on the console
3) ends after some time his X session, to come back to the console

he will see a different behavior compared to 2.6.23 exiting his X session and (I think) believes to have a bug in the X package.

Andrew Morton describes the problem here, too:
http://lkml.org/lkml/2006/11/23/15
http://lkml.org/lkml/2006/11/23/19

Quoted 8 lines

Am I wrong in the assumption, but should one not accept an unchanged behavior with or without capabilities in the kernel regarding the behavior of applications, when he is not actually using (by not setting the xattr capability) capabilities with this application?

If I'm wrong, maybe a warning or hint should be given that one has to grant cap_kill to xinit to come back to the console if the X session was started by xinit.

Thanks - yes, I see (I tend to get lost in my testruns). So we're back to trying to do the fix I was trying to do along with the SIGCONT fix a few weeks ago.

The problem is that when you run a setuid binary, its pP and pE are fully raised. The following patch fixes it for me. Chris, does it fix your problem?

Quoted 7 lines
Quoted 5 lines

Yes, this patch fixes it for me, too.

Thanks,

Chris

Thanks for the valuable testing and bugreports, Chris.

You're welcome!
Chris

Quoted 5 lines

If I don't hear any counter-arguments I'll post the patch in its own thread tomorrow.

-serge

--------------------
Chris Friedhoff
[email protected]

#12 lore raw

Re: Posix file capabilities in 2.6.24rc2; now 2.6.24-rc3

In reply to Serge E. Hallyn

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 4 lines · Serge E. Hallyn

The problem is that when you run a setuid binary, its pP and pE are fully raised. The following patch fixes it for me. Chris, does it fix your problem? Andrew, am I again confusing myself and doing something unsafe?

I think this is yet another example of the fragile mess that is UID emulation with capabilities. Your patch is an example of privilege escalation - luser can kill a more-capable process. In the kill CONT case we reached the opposite conclusion to this one. As was the case then, I didn't disagree then :*). If it meets folk's expectations, then this is probably a good patch...

security/commoncap.c int cap_task_kill(struct task_struct *p, struct siginfo *info, · line 548

Quoted 9 lines
security/commoncap.c +3 -0
@@ -543,6 +543,9 @@ int cap_task_kill(struct task_struct *p, struct siginfo *info,
 	if (capable(CAP_KILL))
 		return 0;
 
+	if (p->euid==0 && p->uid==current->uid)
+		return 0;
+

Its late and I'm obviously tired, but is there any reason not to simply use:

if (p->uid == current->uid)

	 return 0;

Whatever the case, could you put the new code closer to the sig == SIGCONT test? The capability tests are at the end of cap_task_kill() and this new check breaks that pattern.

Cheers

Andrew

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFHRTLlQheEq9QabfIRAt/hAKCJgj2kbuyAWI486LOwwDLdkbcpoQCfQdrQ J+bcvi+9pGTodFn42PsHJHA=
=cXaG
-----END PGP SIGNATURE-----