- 24 Jul, 2014 1 commit
-
-
Eric Paris authored
This is effectively a revert of 7b9a7ec5 plus fixing it a different way... We found, when trying to run an application from an application which had dropped privs that the kernel does security checks on undefined capability bits. This was ESPECIALLY difficult to debug as those undefined bits are hidden from /proc/$PID/status. Consider a root application which drops all capabilities from ALL 4 capability sets. We assume, since the application is going to set eff/perm/inh from an array that it will clear not only the defined caps less than CAP_LAST_CAP, but also the higher 28ish bits which are undefined future capabilities. The BSET gets cleared differently. Instead it is cleared one bit at a time. The problem here is that in security/commoncap.c::cap_task_prctl() we actually check the validity of a capability being read. So any task which attempts to 'read all things set in bset' followed by 'unset all things set in bset' will not even attempt to unset the undefined bits higher than CAP_LAST_CAP. So the 'parent' will look something like: CapInh: 0000000000000000 CapPrm: 0000000000000000 CapEff: 0000000000000000 CapBnd: ffffffc000000000 All of this 'should' be fine. Given that these are undefined bits that aren't supposed to have anything to do with permissions. But they do... So lets now consider a task which cleared the eff/perm/inh completely and cleared all of the valid caps in the bset (but not the invalid caps it couldn't read out of the kernel). We know that this is exactly what the libcap-ng library does and what the go capabilities library does. They both leave you in that above situation if you try to clear all of you capapabilities from all 4 sets. If that root task calls execve() the child task will pick up all caps not blocked by the bset. The bset however does not block bits higher than CAP_LAST_CAP. So now the child task has bits in eff which are not in the parent. These are 'meaningless' undefined bits, but still bits which the parent doesn't have. The problem is now in cred_cap_issubset() (or any operation which does a subset test) as the child, while a subset for valid cap bits, is not a subset for invalid cap bits! So now we set durring commit creds that the child is not dumpable. Given it is 'more priv' than its parent. It also means the parent cannot ptrace the child and other stupidity. The solution here: 1) stop hiding capability bits in status This makes debugging easier! 2) stop giving any task undefined capability bits. it's simple, it you don't put those invalid bits in CAP_FULL_SET you won't get them in init and you won't get them in any other task either. This fixes the cap_issubset() tests and resulting fallout (which made the init task in a docker container untraceable among other things) 3) mask out undefined bits when sys_capset() is called as it might use ~0, ~0 to denote 'all capabilities' for backward/forward compatibility. This lets 'capsh --caps="all=eip" -- -c /bin/bash' run. 4) mask out undefined bit when we read a file capability off of disk as again likely all bits are set in the xattr for forward/backward compatibility. This lets 'setcap all+pe /bin/bash; /bin/bash' run Signed-off-by:
Eric Paris <eparis@redhat.com> Reviewed-by:
Kees Cook <keescook@chromium.org> Cc: Andrew Vagin <avagin@openvz.org> Cc: Andrew G. Morgan <morgan@kernel.org> Cc: Serge E. Hallyn <serge.hallyn@canonical.com> Cc: Kees Cook <keescook@chromium.org> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Dan Walsh <dwalsh@redhat.com> Cc: stable@vger.kernel.org Signed-off-by:
James Morris <james.l.morris@oracle.com>
-
- 06 Jun, 2014 1 commit
-
-
Paul McQuade authored
Use #include <linux/uaccess.h> instead of <asm/uaccess.h> Use #include <linux/types.h> instead of <asm/types.h> Signed-off-by:
Paul McQuade <paulmcquad@gmail.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
- 24 Apr, 2014 1 commit
-
-
Eric W. Biederman authored
It is possible by passing a netlink socket to a more privileged executable and then to fool that executable into writing to the socket data that happens to be valid netlink message to do something that privileged executable did not intend to do. To keep this from happening replace bare capable and ns_capable calls with netlink_capable, netlink_net_calls and netlink_ns_capable calls. Which act the same as the previous calls except they verify that the opener of the socket had the desired permissions as well. Reported-by:
Andy Lutomirski <luto@amacapital.net> Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
- 23 Apr, 2014 3 commits
-
-
Richard Guy Briggs authored
Test first to see if there are any userspace multicast listeners bound to the socket before starting the multicast send work. Signed-off-by:
Richard Guy Briggs <rgb@redhat.com> Signed-off-by:
-
Richard Guy Briggs authored
Add a netlink multicast socket with one group to kaudit for "best-effort" delivery to read-only userspace clients such as systemd, in addition to the existing bidirectional unicast auditd userspace client. Currently, auditd is intended to use the CAP_AUDIT_CONTROL and CAP_AUDIT_WRITE capabilities, but actually uses CAP_NET_ADMIN. The CAP_AUDIT_READ capability is added for use by read-only AUDIT_NLGRP_READLOG netlink multicast group clients to the kaudit subsystem. This will safely give access to services such as systemd to consume audit logs while ensuring write access remains restricted for integrity. Signed-off-by:
-
Richard Guy Briggs authored
Register a netlink per-protocol bind fuction for audit to check userspace process capabilities before allowing a multicast group connection. Signed-off-by:
-
- 31 Mar, 2014 2 commits
-
-
Eric Paris authored
It its possible to configure your PAM stack to refuse login if audit messages (about the login) were unable to be sent. This is common in many distros and thus normal configuration of many containers. The PAM modules determine if audit is enabled/disabled in the kernel based on the return value from sending an audit message on the netlink socket. If userspace gets back ECONNREFUSED it believes audit is disabled in the kernel. If it gets any other error else it refuses to let the login proceed. Just about ever since the introduction of namespaces the kernel audit subsystem has returned EPERM if the task sending a message was not in the init user or pid namespace. So many forms of containers have never worked if audit was enabled in the kernel. BUT if the container was not in net_init then the kernel network code would send ECONNREFUSED (instead of the audit code sending EPERM). Thus by pure accident/dumb luck/bug if an admin configured the PAM stack to reject all logins that didn't talk to audit, but then ran the login untility in the non-init_net namespace, it would work!! Clearly this was a bug, but it is a bug some people expected. With the introduction of network namespace support in 3.14-rc1 the two bugs stopped cancelling each other out. Now, containers in the non-init_net namespace refused to let users log in (just like PAM was configfured!) Obviously some people were not happy that what used to let users log in, now didn't! This fix is kinda hacky. We return ECONNREFUSED for all non-init relevant namespaces. That means that not only will the old broken non-init_net setups continue to work, now the broken non-init_pid or non-init_user setups will 'work'. They don't really work, since audit isn't logging things. But it's what most users want. In 3.15 we should have patches to support not only the non-init_net (3.14) namespace but also the non-init_pid and non-init_user namespace. So all will be right in the world. This just opens the doors wide open on 3.14 and hopefully makes users happy, if not the audit system... Reported-by:
Andre Tomt <andre@tomt.net> Reported-by:
Adam Richter <adam_richter2004@yahoo.com> Signed-off-by:
-
Eric Paris authored
It its possible to configure your PAM stack to refuse login if audit messages (about the login) were unable to be sent. This is common in many distros and thus normal configuration of many containers. The PAM modules determine if audit is enabled/disabled in the kernel based on the return value from sending an audit message on the netlink socket. If userspace gets back ECONNREFUSED it believes audit is disabled in the kernel. If it gets any other error else it refuses to let the login proceed. Just about ever since the introduction of namespaces the kernel audit subsystem has returned EPERM if the task sending a message was not in the init user or pid namespace. So many forms of containers have never worked if audit was enabled in the kernel. BUT if the container was not in net_init then the kernel network code would send ECONNREFUSED (instead of the audit code sending EPERM). Thus by pure accident/dumb luck/bug if an admin configured the PAM stack to reject all logins that didn't talk to audit, but then ran the login untility in the non-init_net namespace, it would work!! Clearly this was a bug, but it is a bug some people expected. With the introduction of network namespace support in 3.14-rc1 the two bugs stopped cancelling each other out. Now, containers in the non-init_net namespace refused to let users log in (just like PAM was configfured!) Obviously some people were not happy that what used to let users log in, now didn't! This fix is kinda hacky. We return ECONNREFUSED for all non-init relevant namespaces. That means that not only will the old broken non-init_net setups continue to work, now the broken non-init_pid or non-init_user setups will 'work'. They don't really work, since audit isn't logging things. But it's what most users want. In 3.15 we should have patches to support not only the non-init_net (3.14) namespace but also the non-init_pid and non-init_user namespace. So all will be right in the world. This just opens the doors wide open on 3.14 and hopefully makes users happy, if not the audit system... Reported-by:
-
- 24 Mar, 2014 1 commit
-
-
Monam Agarwal authored
This patch replaces rcu_assign_pointer(x, NULL) with RCU_INIT_POINTER(x, NULL) The rcu_assign_pointer() ensures that the initialization of a structure is carried out before storing a pointer to that structure. And in the case of the NULL pointer, there is no structure to initialize. So, rcu_assign_pointer(p, NULL) can be safely converted to RCU_INIT_POINTER(p, NULL) Signed-off-by:
Monam Agarwal <monamagarwal123@gmail.com> Signed-off-by:
-
- 20 Mar, 2014 6 commits
-
-
Josh Boyer authored
Calling audit_log_lost with a \n in the format string leads to extra newlines in dmesg. That function will eventually call audit_panic which uses pr_err with an explicit \n included. Just make these calls match the others that lack \n. Reported-by:
Jonathan Kamens <jik@kamens.brookline.ma.us> Signed-off-by:
Josh Boyer <jwboyer@fedoraproject.org> Signed-off-by:
-
Richard Guy Briggs authored
Still only permit the audit logging daemon and control to operate from the initial PID namespace, but allow processes to log from another PID namespace. Cc: "Eric W. Biederman" <ebiederm@xmission.com> (informed by ebiederman's c776b5d2) Signed-off-by: -
Richard Guy Briggs authored
Store and log all PIDs with reference to the initial PID namespace and use the access functions task_pid_nr() and task_tgid_nr() for task->pid and task->tgid. Cc: "Eric W. Biederman" <ebiederm@xmission.com> (informed by ebiederman's c776b5d2) Signed-off-by: -
Richard Guy Briggs authored
sys_getppid() returns the parent pid of the current process in its own pid namespace. Since audit filters are based in the init pid namespace, a process could avoid a filter or trigger an unintended one by being in an alternate pid namespace or log meaningless information. Switch to task_ppid_nr() for PPIDs to anchor all audit filters in the init_pid_ns. (informed by ebiederman's 6c621b7e) Cc: stable@vger.kernel.org Cc: Eric W. Biederman <ebiederm@xmission.com> Signed-off-by: -
Eric W. Biederman authored
In perverse cases of file descriptor passing the current network namespace of a process and the network namespace of a socket used by that socket may differ. Therefore use the network namespace of the appropiate socket to ensure replies always go to the appropiate socket. Signed-off-by:
-
Eric W. Biederman authored
While reading through 3.14-rc1 I found a pretty siginficant mishandling of network namespaces in the recent audit changes. In struct audit_netlink_list and audit_reply add a reference to the network namespace of the caller and remove the userspace pid of the caller. This cleanly remembers the callers network namespace, and removes a huge class of races and nasty failure modes that can occur when attempting to relook up the callers network namespace from a pid_t (including the caller's network namespace changing, pid wraparound, and the pid simply not being present). Signed-off-by:
-
- 08 Mar, 2014 1 commit
-
-
Eric W. Biederman authored
The kbuild test robot reported: > tree: git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace.git for-next > head: 6f285b19 > commit: 6f285b19 [2/2] audit: Send replies in the proper network namespace. > reproduce: make htmldocs > > >> Warning(kernel/audit.c:575): No description found for parameter 'request_skb' > >> Warning(kernel/audit.c:575): Excess function parameter 'portid' description in 'audit_send_reply' > >> Warning(kernel/auditfilter.c:1074): No description found for parameter 'request_skb' > >> Warning(kernel/auditfilter.c:1074): Excess function parameter 'portid' description in 'audit_list_rules_s Which was caused by my failure to update the kdoc annotations when I updated the functions. Fix that small oversight now. Signed-off-by:
-
- 01 Mar, 2014 1 commit
-
-
Eric W. Biederman authored
In perverse cases of file descriptor passing the current network namespace of a process and the network namespace of a socket used by that socket may differ. Therefore use the network namespace of the appropiate socket to ensure replies always go to the appropiate socket. Signed-off-by:
-
- 28 Feb, 2014 1 commit
-
-
Eric W. Biederman authored
In struct audit_netlink_list and audit_reply add a reference to the network namespace of the caller and remove the userspace pid of the caller. This cleanly remembers the callers network namespace, and removes a huge class of races and nasty failure modes that can occur when attempting to relook up the callers network namespace from a pid_t (including the caller's network namespace changing, pid wraparound, and the pid simply not being present). Signed-off-by:
-
- 17 Jan, 2014 2 commits
-
-
Richard Guy Briggs authored
Fixup caught by checkpatch. Signed-off-by:
-
Eric Paris authored
A message about creating the audit socket might be fine at startup, but a pr_info for every single network namespace created on a system isn't useful. Signed-off-by:
-
- 14 Jan, 2014 20 commits
-
-
Joe Perches authored
The equivalent uapi struct uses __u32 so make the kernel uses u32 too. This can prevent some oddities where the limit is logged/emitted as a negative value. Convert kstrtol to kstrtouint to disallow negative values. Signed-off-by:Joe Perches <joe@perches.com> [eparis: do not remove static from audit_default declaration]
-
Joe Perches authored
Add pr_fmt to prefix "audit: " to output Convert printk(KERN_<LEVEL> to pr_<level> Coalesce formats Use pr_cont Move a brace after switch Signed-off-by: -
Joe Perches authored
Using the generic kernel function causes the object size to increase with gcc 4.8.1. $ size kernel/audit.o* text data bss dec hex filename 18577 6079 8436 33092 8144 kernel/audit.o.new 18579 6015 8420 33014 80f6 kernel/audit.o.old Unsigned...
-
Eric Paris authored
An admin is likely to want to see old and new values next to each other. Putting all of the old values followed by all of the new values is just hard to read as a human. Signed-off-by: -
Eric Paris authored
We can simplify the AUDIT_TTY_SET code to only grab the spin_lock one time. We need to determine if the new values are valid and if so, set the new values at the same time we grab the old onces. While we are here get rid of 'res' and just use err. Signed-off-by: -
Eric Paris authored
If userspace specified that it was setting values via the mask we do not need a second check to see if they also set the version field high enough to understand those values. (clearly if they set the mask they knew those values). Signed-off-by: -
Eric Paris authored
Give names to the audit versions. Just something for a userspace programmer to know what the version provides. Signed-off-by: -
Eric Paris authored
We had some craziness with signed to unsigned long casting which appears wholely unnecessary. Just use signed long. Even though 2 values of the math equation are unsigned longs the result is expected to be a signed long. So why keep casting the result to signed long? Just make it signed long and use it. We also remove the needless "timeout" variable. We already have the stack "sleep_time" variable. Just use that... Signed-off-by: -
Richard Guy Briggs authored
Add task information to the log when changing a feature state. Signed-off-by: -
Gao feng authored
NETLINK_CB(skb).sk is the socket of user space process, netlink_unicast in kauditd_send_skb wants the kernel side socket. Since the sk_state of audit netlink socket is not NETLINK_CONNECTED, so the netlink_getsockbyportid doesn't return -ECONNREFUSED. And the socket of userspace process can be released anytime, so the audit_sock may point to invalid socket. this patch sets the audit_sock to the kernel side audit netlink socket. Signed-off-by:
Gao feng <gaofeng@cn.fujitsu.com> Acked-by:
-
Gao feng authored
print the error message and then return -ENOMEM. Signed-off-by:
-
Richard Guy Briggs authored
An error on an AUDIT_NEVER rule disabled logging on that rule. On error on AUDIT_NEVER rules, log. Signed-off-by:
-
Toshiyuki Okajima authored
The backlog cannot be consumed when audit_log_start is running on auditd even if audit_log_start calls wait_for_auditd to consume it. The situation is the deadlock because only auditd can consume the backlog. If the other process needs to send the backlog, it can be also stopped by the deadlock. So, audit_log_start running on auditd should not stop. You can see the deadlock with the following reproducer: # auditctl -a exit,always -S all # reboot Signed-off-by:
Toshiyuki Okajima <toshi.okajima@jp.fujitsu.com> Reviewed-by: gaofeng@cn.fujitsu.com Signed-off-by:
-
Richard Guy Briggs authored
We do not need to hold the audit_cmd_mutex for this family of cases. The possible exception to this is the call to audit_filter_user(), so drop the lock immediately after. To help in fixing the race we are trying to avoid, make sure that nothing called by audit_filter_user() calls audit_log_start(). In particular, watch out for *_audit_rule_match(). This fix will take care of systemd and anything USING audit. It still means that we could race with something configuring audit and auditd shutting down. Signed-off-by:
-
Eric Paris authored
Right now the sessionid value in the kernel is a combination of u32, int, and unsigned int. Just use unsigned int throughout. Signed-off-by:
-
Paul Davies C authored
Currently when the coredump signals are logged by the audit system, the actual path to the executable is not logged. Without details of exe, the system admin may not have an exact idea on what program failed. This patch changes the audit_log_task() so that the path to the exe is also logged. This was copied from audit_log_task_info() and the latter enhanced to avoid disappearing text fields. Signed-off-by:
Paul Davies C <pauldaviesc@gmail.com> Signed-off-by:
-
Richard Guy Briggs authored
There have been reports of auditd restarts resulting in kaudit not being able to find a newly registered auditd. It results in reports such as: kernel: [ 2077.233573] audit: *NO* daemon at audit_pid=1614 kernel: [ 2077.234712] audit: audit_lost=97 audit_rate_limit=0 audit_backlog_limit=320 kernel: [ 2077.234718] audit: auditd disappeared (previously mis-spelled "dissapeared") One possible cause is a race between the shutdown of an older auditd and a newer one. If the newer one sets the daemon pid to itself in kauditd before the older one has cleared the daemon pid, the newer daemon pid will be erased. This could be caused by an automated system, or by manual intervention, but in either case, there is no use in having the older daemon clear the daemon pid reference since its old pid is no longer being referenced. This patch will prevent that specific case, returning an error of EACCES. The case for preventing a newer auditd from registering itself if there is an existing auditd is a more difficult case that is beyond the scope of this patch. Signed-off-by:
-
Richard Guy Briggs authored
audit_receive_msg() needlessly contained a fallthrough case that called audit_receive_filter(), containing no common code between the cases. Separate them to make the logic clearer. Refactor AUDIT_LIST_RULES, AUDIT_ADD_RULE, AUDIT_DEL_RULE cases to create audit_rule_change(), audit_list_rules_send() functions. This should not functionally change the logic. Signed-off-by:
-
Richard Guy Briggs authored
Log transition of config changes when AUDIT_TTY_SET is called, including both enabled and log_passwd values now in the struct. Signed-off-by:
-
Richard Guy Briggs authored
kauditd_send_skb is called after audit_pid was checked to be non-zero. However, it can be set to 0 due to auditd exiting while kauditd_send_skb is still executed and this can result in a spurious warning about missing auditd. Re-check audit_pid before printing the message. Signed-off-by:
Mateusz Guzik <mguzik@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: linux-kernel@vger.kernel.org Acked-by:
-
