summaryrefslogtreecommitdiffstats
path: root/Documentation
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2021-11-13 10:27:50 -0800
committerLinus Torvalds <torvalds@linux-foundation.org>2021-11-13 10:27:50 -0800
commit0a90729278ae7b31084d2d436b0eee4d83b11506 (patch)
tree9b662a87d00e145dbce8146f6dc9202abce33db4 /Documentation
parent7c3737c706073792133deeefae33ab17fd06e0c2 (diff)
parent32a370abf12f82c8383e430c21365f5355d8b288 (diff)
downloadlinux-0a90729278ae7b31084d2d436b0eee4d83b11506.tar.bz2
Merge tag 'selinux-pr-20211112' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux
Pull selinux fixes from Paul Moore: "Unfortunately I need to request a revert for two LSM/SELinux patches that came in via the network tree. The two patches in question add a new SCTP/LSM hook as well as an SELinux implementation of that LSM hook. The short version of "why?" is in the commit description of the revert patch, but I'll copy-n-paste the important bits below to save some time for the curious: ... Unfortunately these two patches were merged without proper review (the Reviewed-by and Tested-by tags from Richard Haines were for previous revisions of these patches that were significantly different) and there are outstanding objections from the SELinux maintainers regarding these patches. Work is currently ongoing to correct the problems identified in the reverted patches, as well as others that have come up during review, but it is unclear at this point in time when that work will be ready for inclusion in the mainline kernel. In the interest of not keeping objectionable code in the kernel for multiple weeks, and potentially a kernel release, we are reverting the two problematic patches. As usual with these things there is plenty of context to go with this and I'll try to do my best to provide that now. This effort started with a report of SCTP client side peel-offs not working correctly with SELinux, Ondrej Mosnacek put forth a patch which he believed properly addressed the problem but upon review by the netdev folks Xin Long described some additional issues and submitted an improved patchset for review. The SELinux folks reviewed Xin Long's initial patchset and suggested some changes which resulted in a second patchset (v2) from Xin Long; this is the patchset that is currently in your tree. Unfortunately this v2 patchset from Xin Long was merged before it had spent even just 24 hours on the mailing lists during the early days of the merge window, a time when many of us were busy doing verification of the newly released v5.15 kernel as well final review and testing of our v5.16 pull requests. Making matters worse, upon reviewing the v2 patchset there were both changes which were found objectionable by SELinux standards as well as additional outstanding SCTP/SELinux interaction problems. At this point we did two things: resumed working on a better fix for the SCTP/SELinux issue(s) - thank you Ondrej - and we asked the networking folks to revert the v2 patchset. The revert request was obviously rejected, but at the time I believed it was just going to be an issue for linux-next; I wasn't expecting something this significant that was merged into the networking tree during the merge window to make it into your tree in the same window, yet as of last night that is exactly what happened. While we continue to try and resolve the SCTP/SELinux problem I am asking once again to revert the v2 patches and not ship the current security_sctp_assoc_established() hook in a v5.16-rcX kernel. If I was confident that we could solve these issues in a week, maybe two, I would refrain from asking for the revert but our current estimate is for a minimum of two weeks for the next patch revision. With the likelihood of additional delays due to normal patch review follow-up and/or holidays it seems to me that the safest course of action is to revert the patch both to try and keep some objectionable code out of a release kernel and limit the chances of any new breakages from such a change. While the SCTP/SELinux code in v5.15 and earlier has problems, they are known problems, and I'd like to try and avoid creating new and different problems while we work to fix things properly. One final thing to mention: Xin Long's v2 patchset consisted of four patches, yet this revert is for only the last two. We see the first two patches as good, reasonable, and not likely to cause an issue. In an attempt to create a cleaner revert patch we suggest leaving the first two patches in the tree as they are currently" * tag 'selinux-pr-20211112' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux: net,lsm,selinux: revert the security_sctp_assoc_established() hook
Diffstat (limited to 'Documentation')
-rw-r--r--Documentation/security/SCTP.rst22
1 files changed, 12 insertions, 10 deletions
diff --git a/Documentation/security/SCTP.rst b/Documentation/security/SCTP.rst
index 406cc68b8808..d5fd6ccc3dcb 100644
--- a/Documentation/security/SCTP.rst
+++ b/Documentation/security/SCTP.rst
@@ -15,7 +15,10 @@ For security module support, three SCTP specific hooks have been implemented::
security_sctp_assoc_request()
security_sctp_bind_connect()
security_sctp_sk_clone()
- security_sctp_assoc_established()
+
+Also the following security hook has been utilised::
+
+ security_inet_conn_established()
The usage of these hooks are described below with the SELinux implementation
described in the `SCTP SELinux Support`_ chapter.
@@ -119,12 +122,11 @@ calls **sctp_peeloff**\(3).
@newsk - pointer to new sock structure.
-security_sctp_assoc_established()
+security_inet_conn_established()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-Called when a COOKIE ACK is received, and the peer secid will be
-saved into ``@asoc->peer_secid`` for client::
+Called when a COOKIE ACK is received::
- @asoc - pointer to sctp association structure.
+ @sk - pointer to sock structure.
@skb - pointer to skbuff of the COOKIE ACK packet.
@@ -132,7 +134,7 @@ Security Hooks used for Association Establishment
-------------------------------------------------
The following diagram shows the use of ``security_sctp_bind_connect()``,
-``security_sctp_assoc_request()``, ``security_sctp_assoc_established()`` when
+``security_sctp_assoc_request()``, ``security_inet_conn_established()`` when
establishing an association.
::
@@ -170,7 +172,7 @@ establishing an association.
<------------------------------------------- COOKIE ACK
| |
sctp_sf_do_5_1E_ca |
- Call security_sctp_assoc_established() |
+ Call security_inet_conn_established() |
to set the peer label. |
| |
| If SCTP_SOCKET_TCP or peeled off
@@ -196,7 +198,7 @@ hooks with the SELinux specifics expanded below::
security_sctp_assoc_request()
security_sctp_bind_connect()
security_sctp_sk_clone()
- security_sctp_assoc_established()
+ security_inet_conn_established()
security_sctp_assoc_request()
@@ -269,12 +271,12 @@ sockets sid and peer sid to that contained in the ``@asoc sid`` and
@newsk - pointer to new sock structure.
-security_sctp_assoc_established()
+security_inet_conn_established()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Called when a COOKIE ACK is received where it sets the connection's peer sid
to that in ``@skb``::
- @asoc - pointer to sctp association structure.
+ @sk - pointer to sock structure.
@skb - pointer to skbuff of the COOKIE ACK packet.