AVC denied for crio on access of tmpfs (modules_object_t) #89
Comments
|
Very strange this looks like you have a file system labelled modules_object_t? This AVC would indicate you were attempting to put a file system object labeled container_file_t on a file system labeled modules_object_t. Do you have any idea what container you are executing when this happens? @stephensmalley Any ideas? |
|
In this case it is a Fluentd container, with some specific mounts to the host: Thought it affects all running containers. Will check. |
|
Since we're not seeing this issue with every cluster deployment (and cannot reproduce this easily) we're a bit dependent on the moments this occurs. modules_objects_t is only defined for: No mounted filesystems with To make it even more obscure: The moment we see the avc denied for crio, we also see a simmilar one for systemd-logind: and for crio A set of files and directories in Restarting the kubelet systemd unit doesn't resolve this issue, rebooting the node does. |
|
Are these file systems being mounted with a context mount? On the host could you do a |
|
Context The tmpfs mount belonging to the above snippet is:
|
|
@stephensmalley Any ideas? |
|
If you don't have any context mounts with modules_object_t and you don't have any fs_use or genfscon rules in policy that specify modules_object_t then this seems like a kernel bug. |
|
@wrabcak FYI? |
|
Very strange. |
|
No |
|
With the risk of jinxing the happy streak: I haven't seen the issue occur in over +50 deployments after upgrading cri-o/k8s to 1.17.4. Thanks for the input! |
pcmoore
pushed a commit
to SELinuxProject/selinux-kernel
that referenced
this issue
Oct 11, 2021
ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
pcmoore
pushed a commit
to SELinuxProject/selinux-kernel
that referenced
this issue
Oct 12, 2021
Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
imaami
pushed a commit
to imaami/linux
that referenced
this issue
Nov 13, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
woodsts
pushed a commit
to woodsts/linux-stable
that referenced
this issue
Nov 17, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
woodsts
pushed a commit
to woodsts/linux-stable
that referenced
this issue
Nov 18, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Whissi
pushed a commit
to Whissi/linux-stable
that referenced
this issue
Nov 18, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
it-is-a-robot
pushed a commit
to openeuler-mirror/kernel
that referenced
this issue
Dec 8, 2021
stable inclusion from stable-5.10.80 commit db04fb4111e6f5cf1518c7a4718ccde1756f909e bugzilla: 185821 https://gitee.com/openeuler/kernel/issues/I4L7CG Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=db04fb4111e6f5cf1518c7a4718ccde1756f909e -------------------------------- commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Chen Jun <chenjun102@huawei.com> Reviewed-by: Weilong Chen <chenweilong@huawei.com> Acked-by: Weilong Chen <chenweilong@huawei.com> Signed-off-by: Chen Jun <chenjun102@huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com>
ammarfaizi2
pushed a commit
to ammarfaizi2/linux-block
that referenced
this issue
Dec 16, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ammarfaizi2
pushed a commit
to ammarfaizi2/linux-block
that referenced
this issue
Dec 16, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
woodsts
pushed a commit
to woodsts/linux-stable
that referenced
this issue
Dec 17, 2021
commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
jpuhlman
pushed a commit
to MontaVista-OpenSourceTechnology/linux-mvista
that referenced
this issue
Dec 21, 2021
Source: Kernel.org MR: 114068 Type: Integration Disposition: Backport from git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable linux-5.4.y ChangeID: b06b1f46306a5fdcb9da8c9cf59e643a357ff6b8 Description: commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Armin Kuster <akuster@mvista.com>
Matombo
pushed a commit
to tuxedocomputers/linux
that referenced
this issue
Jan 26, 2022
BugLink: https://bugs.launchpad.net/bugs/1954337 commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Kelsey Skunberg <kelsey.skunberg@canonical.com>
delphix-devops-bot
pushed a commit
to delphix/linux-kernel-oracle
that referenced
this issue
Feb 2, 2022
BugLink: https://bugs.launchpad.net/bugs/1954533 commit cbfcd13be5cb2a07868afe67520ed181956579a7 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Kaz205
pushed a commit
to Kaz205/renoir
that referenced
this issue
Feb 3, 2022
Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
ArtiicPanda
pushed a commit
to ArtiicPanda/oneplus-sm4350
that referenced
this issue
Feb 12, 2022
commit cbfcd13be5cb2a07868afe67520ed181956579a7 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kaz205
pushed a commit
to Kaz205/renoir
that referenced
this issue
Feb 13, 2022
Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
Kaz205
pushed a commit
to Kaz205/renoir
that referenced
this issue
Feb 13, 2022
Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
Kaz205
pushed a commit
to Kaz205/renoir
that referenced
this issue
Feb 16, 2022
Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
Kaz205
pushed a commit
to Kaz205/renoir
that referenced
this issue
Feb 16, 2022
Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
dlpx-writer bot
pushed a commit
to delphix/linux-kernel-aws
that referenced
this issue
Mar 25, 2022
BugLink: https://bugs.launchpad.net/bugs/1957987 commit cbfcd13be5cb2a07868afe67520ed181956579a7 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
hisenyiu2015
pushed a commit
to hisenyiu2015/msm-4.14
that referenced
this issue
Mar 29, 2022
BugLink: https://bugs.launchpad.net/bugs/1951822 commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Apr 2, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
May 10, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Jul 8, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Jul 8, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Aug 23, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Aug 23, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Nov 5, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Nov 5, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Nov 5, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Nov 5, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
nvidia-bfigg
pushed a commit
to NVIDIA-BaseOS-6/linux-nvidia-5.19
that referenced
this issue
Nov 15, 2022
BugLink: https://bugs.launchpad.net/bugs/1951822 commit cbfcd13 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
Reinazhard
pushed a commit
to Reinazhard/android_kernel_xiaomi_whyred
that referenced
this issue
Dec 6, 2022
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Rohail33
pushed a commit
to Rohail33/Realking_xiaomi_xaga
that referenced
this issue
Dec 31, 2022
commit cbfcd13be5cb2a07868afe67520ed181956579a7 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
meydiwahendra
pushed a commit
to meydiwahendra/android_kernel_xiaomi_whyred
that referenced
this issue
Sep 9, 2023
Current code contains a lot of racy patterns when converting an
ocontext's context structure to an SID. This is being done in a "lazy"
fashion, such that the SID is looked up in the SID table only when it's
first needed and then cached in the "sid" field of the ocontext
structure. However, this is done without any locking or memory barriers
and is thus unsafe.
Between commits 24ed7fdae669 ("selinux: use separate table for initial
SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash
table"), this race condition lead to an actual observable bug, because a
pointer to the shared sid field was passed directly to
sidtab_context_to_sid(), which was using this location to also store an
intermediate value, which could have been read by other threads and
interpreted as an SID. In practice this caused e.g. new mounts to get a
wrong (seemingly random) filesystem context, leading to strange denials.
This bug has been spotted in the wild at least twice, see [1] and [2].
Fix the race condition by making all the racy functions use a common
helper that ensures the ocontext::sid accesses are made safely using the
appropriate SMP constructs.
Note that security_netif_sid() was populating the sid field of both
contexts stored in the ocontext, but only the first one was actually
used. The SELinux wiki's documentation on the "netifcon" policy
statement [3] suggests that using only the first context is intentional.
I kept only the handling of the first context here, as there is really
no point in doing the SID lookup for the unused one.
I wasn't able to reproduce the bug mentioned above on any kernel that
includes commit 66f8e2f03c02, even though it has been reported that the
issue occurs with that commit, too, just less frequently. Thus, I wasn't
able to verify that this patch fixes the issue, but it makes sense to
avoid the race condition regardless.
[1] containers/container-selinux#89
[2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/
[3] https://selinuxproject.org/page/NetworkStatements#netifcon
Cc: stable@vger.kernel.org
Cc: Xinjie Zheng <xinjie@google.com>
Reported-by: Sujithra Periasamy <sujithra@google.com>
Fixes: 1da177e ("Linux-2.6.12-rc2")
Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Signed-off-by: meydiwahendra <meydiwahendra@gmail.com>
Dheeraj3031A
pushed a commit
to Dheeraj3031A/kernel_oplus_RMX3461
that referenced
this issue
Jan 31, 2024
commit cbfcd13be5cb2a07868afe67520ed181956579a7 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> [vijayb: Backport contextual differences are due to v5.10 RCU related changes are not in 5.4] Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
electimon
pushed a commit
to moto-common/android_kernel_motorola_msm-5.4
that referenced
this issue
Jun 6, 2024
commit cbfcd13be5cb2a07868afe67520ed181956579a7 upstream. Current code contains a lot of racy patterns when converting an ocontext's context structure to an SID. This is being done in a "lazy" fashion, such that the SID is looked up in the SID table only when it's first needed and then cached in the "sid" field of the ocontext structure. However, this is done without any locking or memory barriers and is thus unsafe. Between commits 24ed7fd ("selinux: use separate table for initial SID lookup") and 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table"), this race condition lead to an actual observable bug, because a pointer to the shared sid field was passed directly to sidtab_context_to_sid(), which was using this location to also store an intermediate value, which could have been read by other threads and interpreted as an SID. In practice this caused e.g. new mounts to get a wrong (seemingly random) filesystem context, leading to strange denials. This bug has been spotted in the wild at least twice, see [1] and [2]. Fix the race condition by making all the racy functions use a common helper that ensures the ocontext::sid accesses are made safely using the appropriate SMP constructs. Note that security_netif_sid() was populating the sid field of both contexts stored in the ocontext, but only the first one was actually used. The SELinux wiki's documentation on the "netifcon" policy statement [3] suggests that using only the first context is intentional. I kept only the handling of the first context here, as there is really no point in doing the SID lookup for the unused one. I wasn't able to reproduce the bug mentioned above on any kernel that includes commit 66f8e2f03c02, even though it has been reported that the issue occurs with that commit, too, just less frequently. Thus, I wasn't able to verify that this patch fixes the issue, but it makes sense to avoid the race condition regardless. [1] containers/container-selinux#89 [2] https://lists.fedoraproject.org/archives/list/selinux@lists.fedoraproject.org/thread/6DMTAMHIOAOEMUAVTULJD45JZU7IBAFM/ [3] https://selinuxproject.org/page/NetworkStatements#netifcon Cc: stable@vger.kernel.org Cc: Xinjie Zheng <xinjie@google.com> Reported-by: Sujithra Periasamy <sujithra@google.com> Fixes: 1da177e ("Linux-2.6.12-rc2") Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Occasionally I run into an AVC denied when running CRI-O 1.14.11 on RHEL7.7. This occurs once every N times I redeploy a host. The issue persists after stop/start-ing
crio.service. This probably excluded issues with the order in which (selinux-policy) RPM's are installed and services are started.The SEalert report from this log-entry is:
I've gathered the log-data after switching SElinux to Permissive.
The text was updated successfully, but these errors were encountered: