Skip to contentSkip to conversation

cap-bound not working in uml

uml-devel Frank v Waveren 3 participants 7 messages Jul 15, 2006 – Jul 18, 2006 User-mode Linux (UML)

lore
Share & tools
lore
#0 lore raw

I was trying to limit some unecessary capabilities in a UML instance with /proc/sys/kernel/cap-bound, but it turned out not to take.

The source of the problem (or at least something a bit of the way up the garden path of the problem) is at security/commoncap.c:140 at the top of cap_bprm_apply_creds(bprm, unsafe):

   void cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe)
   {
           /* Derived from fs/exec.c:compute_creds. */
           kernel_cap_t new_permitted, working;
           new_permitted = cap_intersect (bprm->cap_permitted, cap_bset);
           working = cap_intersect (bprm->cap_inheritable,
                                    current->cap_inheritable);
           new_permitted = cap_combine (new_permitted, working);
           ...

Here the new permitted set gets limited to the bits in cap_bset, which is as it should be, but then the intersection of the of the current and exec inheritable masks get added to that set, whereas as I understand it, cap_bset should always be the bounding set.

I've tried commenting out that bit and everything worked as I'd hoped (I haven't done extensive testing, but bounding the caps worked, as did suids and such).

That doesn't explain why it works with those lines left in on a non-UML kernel though, so I assume I'm missing something fundamental.

(My guest kernel is
Linux version 2.6.16.24 ([email protected]) (gcc version 4.0.3 20051201 (prerelease) (Debian 4.0.2-5)) #3 Sat Jul 15 16:54:20 CEST 2006
, should it matter)

Signature · 3 lines
Frank v Waveren                                  Key fingerprint: BDD7 D61E
[email protected]                                              5D39 CF05 4BFC F57A
Public key: hkp://wwwkeys.pgp.net/468D62C8              FA00 7D51 468D 62C8

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#1 lore raw
Quoted 2 lines · Frank v Waveren

I was trying to limit some unecessary capabilities in a UML instance with /proc/sys/kernel/cap-bound, but it turned out not to take.

To remove capabilities from the whole system (i.e. all processes) the recommended way wasn't to use lcap (or a similar program bundled with libcap)?

Quoted 31 lines

The source of the problem (or at least something a bit of the way up the garden path of the problem) is at security/commoncap.c:140 at the top of cap_bprm_apply_creds(bprm, unsafe):

   void cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe)
   {
           /* Derived from fs/exec.c:compute_creds. */
           kernel_cap_t new_permitted, working;
           new_permitted = cap_intersect (bprm->cap_permitted, cap_bset);
           working = cap_intersect (bprm->cap_inheritable,
                                    current->cap_inheritable);
           new_permitted = cap_combine (new_permitted, working);
           ...

Here the new permitted set gets limited to the bits in cap_bset, which is as it should be, but then the intersection of the of the current and exec inheritable masks get added to that set, whereas as I understand it, cap_bset should always be the bounding set.

I've tried commenting out that bit and everything worked as I'd hoped (I haven't done extensive testing, but bounding the caps worked, as did suids and such).

That doesn't explain why it works with those lines left in on a non-UML kernel though, so I assume I'm missing something fundamental.

(My guest kernel is
Linux version 2.6.16.24 ([email protected]) (gcc version 4.0.3 20051201 (prerelease) (Debian 4.0.2-5)) #3 Sat Jul 15 16:54:20 CEST 2006
, should it matter)

Signature · 17 lines
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade
http://www.user-mode-linux.org/~blaisorblade
Chiacchiera con i tuoi amici in tempo reale! 
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com 



-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
#2 lore raw
Quoted 7 lines · Blaisorblade
Quoted 2 lines · Frank v Waveren

I was trying to limit some unecessary capabilities in a UML instance with /proc/sys/kernel/cap-bound, but it turned out not to take.

To remove capabilities from the whole system (i.e. all processes) the recommended way wasn't to use lcap (or a similar program bundled with libcap)?

Yup, lcap is just an interface to /proc/sys/kernel/cap-bound.

Quoted 31 lines
Quoted 31 lines

The source of the problem (or at least something a bit of the way up the garden path of the problem) is at security/commoncap.c:140 at the top of cap_bprm_apply_creds(bprm, unsafe):

   void cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe)
   {
           /* Derived from fs/exec.c:compute_creds. */
           kernel_cap_t new_permitted, working;
           new_permitted = cap_intersect (bprm->cap_permitted, cap_bset);
           working = cap_intersect (bprm->cap_inheritable,
                                    current->cap_inheritable);
           new_permitted = cap_combine (new_permitted, working);
           ...

Here the new permitted set gets limited to the bits in cap_bset, which is as it should be, but then the intersection of the of the current and exec inheritable masks get added to that set, whereas as I understand it, cap_bset should always be the bounding set.

I've tried commenting out that bit and everything worked as I'd hoped (I haven't done extensive testing, but bounding the caps worked, as did suids and such).

That doesn't explain why it works with those lines left in on a non-UML kernel though, so I assume I'm missing something fundamental.

(My guest kernel is
Linux version 2.6.16.24 ([email protected]) (gcc version 4.0.3 20051201 (prerelease) (Debian 4.0.2-5)) #3 Sat Jul 15 16:54:20 CEST 2006
, should it matter)

Signature · 3 lines
Frank v Waveren                                  Key fingerprint: BDD7 D61E
[email protected]                                              5D39 CF05 4BFC F57A
Public key: hkp://wwwkeys.pgp.net/468D62C8              FA00 7D51 468D 62C8

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#3 lore raw
Quoted 10 lines · Frank v Waveren
Quoted 7 lines · Blaisorblade
Quoted 2 lines · Frank v Waveren

I was trying to limit some unecessary capabilities in a UML instance with /proc/sys/kernel/cap-bound, but it turned out not to take.

To remove capabilities from the whole system (i.e. all processes) the recommended way wasn't to use lcap (or a similar program bundled with libcap)?

Yup, lcap is just an interface to /proc/sys/kernel/cap-bound.

I remember maybe some differences UML specific capabilities code in arch/um (it used to clear some capabilities, i.e. /dev/mem or /dev/kmem access). Check for that...

Quoted 31 lines
Quoted 31 lines
Quoted 31 lines

The source of the problem (or at least something a bit of the way up the garden path of the problem) is at security/commoncap.c:140 at the top of cap_bprm_apply_creds(bprm, unsafe):

   void cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe)
   {
           /* Derived from fs/exec.c:compute_creds. */
           kernel_cap_t new_permitted, working;
           new_permitted = cap_intersect (bprm->cap_permitted,
cap_bset); working = cap_intersect (bprm->cap_inheritable,
                                    current->cap_inheritable);
           new_permitted = cap_combine (new_permitted, working);
           ...

Here the new permitted set gets limited to the bits in cap_bset, which is as it should be, but then the intersection of the of the current and exec inheritable masks get added to that set, whereas as I understand it, cap_bset should always be the bounding set.

I've tried commenting out that bit and everything worked as I'd hoped (I haven't done extensive testing, but bounding the caps worked, as did suids and such).

That doesn't explain why it works with those lines left in on a non-UML kernel though, so I assume I'm missing something fundamental.

(My guest kernel is
Linux version 2.6.16.24 ([email protected]) (gcc version 4.0.3
20051201 (prerelease) (Debian 4.0.2-5)) #3 Sat Jul 15 16:54:20 CEST 2006 , should it matter)

Signature · 17 lines
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade
http://www.user-mode-linux.org/~blaisorblade
Chiacchiera con i tuoi amici in tempo reale! 
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com 



-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
#4 lore raw
Quoted 10 lines · Blaisorblade
Quoted 9 lines · Frank v Waveren
Quoted 7 lines · Blaisorblade
Quoted 2 lines · Frank v Waveren

I was trying to limit some unecessary capabilities in a UML instance with /proc/sys/kernel/cap-bound, but it turned out not to take.

To remove capabilities from the whole system (i.e. all processes) the recommended way wasn't to use lcap (or a similar program bundled with libcap)?

Quoted 5 lines
Quoted 1 line

Yup, lcap is just an interface to /proc/sys/kernel/cap-bound.

I remember maybe some differences UML specific capabilities code in arch/um (it used to clear some capabilities, i.e. /dev/mem or /dev/kmem access). Check for that...

There used to be (in the tt mode jail support), but it's all gone now (leaving some undefined prototypes in the headers about jail mode and such by the way).

Oh well, I'll have to go poke in a live non-UML kernel and see what's different, I'll keep the list posted.

Signature · 3 lines
Frank v Waveren                                  Key fingerprint: BDD7 D61E
[email protected]                                              5D39 CF05 4BFC F57A
Public key: hkp://wwwkeys.pgp.net/468D62C8              FA00 7D51 468D 62C8

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#5 lore raw
In reply to Frank v Waveren

The problem turned out to be caused in userspace, for some odd reason the init on the gentoo disk image I'd nabbed did something that set its inheritable mask to ~0. Not sure if its doing it on purpose, and I still don't quite understand the code I quoted below, but it's not relevant to UML, so I'll ask about it elsewhere. Sorry for the noise.

Quoted 37 lines · Frank v Waveren
Quoted 7 lines · Blaisorblade
Quoted 2 lines · Frank v Waveren

I was trying to limit some unecessary capabilities in a UML instance with /proc/sys/kernel/cap-bound, but it turned out not to take.

To remove capabilities from the whole system (i.e. all processes) the recommended way wasn't to use lcap (or a similar program bundled with libcap)?

Yup, lcap is just an interface to /proc/sys/kernel/cap-bound.

Quoted 26 lines
Quoted 26 lines

The source of the problem (or at least something a bit of the way up the garden path of the problem) is at security/commoncap.c:140 at the top of cap_bprm_apply_creds(bprm, unsafe):

   void cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe)
   {
           /* Derived from fs/exec.c:compute_creds. */
           kernel_cap_t new_permitted, working;
           new_permitted = cap_intersect (bprm->cap_permitted, cap_bset);
           working = cap_intersect (bprm->cap_inheritable,
                                    current->cap_inheritable);
           new_permitted = cap_combine (new_permitted, working);
           ...

Here the new permitted set gets limited to the bits in cap_bset, which is as it should be, but then the intersection of the of the current and exec inheritable masks get added to that set, whereas as I understand it, cap_bset should always be the bounding set.

I've tried commenting out that bit and everything worked as I'd hoped (I haven't done extensive testing, but bounding the caps worked, as did suids and such).

That doesn't explain why it works with those lines left in on a non-UML kernel though, so I assume I'm missing something fundamental.

Signature · 3 lines
Frank v Waveren                                  Key fingerprint: BDD7 D61E
[email protected]                                              5D39 CF05 4BFC F57A
Public key: hkp://wwwkeys.pgp.net/468D62C8              FA00 7D51 468D 62C8

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#6 lore raw

Oh, and I should just add that UML is a great pleasure to debug compared to kernel-mode-linux. Thanks for the wonderful work everyone!

Signature · 3 lines
Frank v Waveren                                  Key fingerprint: BDD7 D61E
[email protected]                                              5D39 CF05 4BFC F57A
Public key: hkp://wwwkeys.pgp.net/468D62C8              FA00 7D51 468D 62C8

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel