linux/kernel/auditsc.c

3043 lines
82 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/* auditsc.c -- System-call auditing support
* Handles all system-call specific auditing features.
*
* Copyright 2003-2004 Red Hat Inc., Durham, North Carolina.
* Copyright 2005 Hewlett-Packard Development Company, L.P.
* Copyright (C) 2005, 2006 IBM Corporation
* All Rights Reserved.
*
* Written by Rickard E. (Rik) Faith <faith@redhat.com>
*
* Many of the ideas implemented here are from Stephen C. Tweedie,
* especially the idea of avoiding a copy by using getname.
*
* The method for actual interception of syscall entry and exit (not in
* this file -- see entry.S) is based on a GPL'd patch written by
* okir@suse.de and Copyright 2003 SuSE Linux AG.
*
* POSIX message queue support added by George Wilson <ltcgcw@us.ibm.com>,
* 2006.
*
* The support of additional filter rules compares (>, <, >=, <=) was
* added by Dustin Kirkland <dustin.kirkland@us.ibm.com>, 2005.
*
* Modified by Amy Griffis <amy.griffis@hp.com> to collect additional
* filesystem information.
*
* Subject and object context labeling support added by <danjones@us.ibm.com>
* and <dustin.kirkland@us.ibm.com> for LSPP certification compliance.
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/init.h>
#include <asm/types.h>
#include <linux/atomic.h>
#include <linux/fs.h>
#include <linux/namei.h>
#include <linux/mm.h>
#include <linux/export.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/slab.h>
#include <linux/mount.h>
#include <linux/socket.h>
#include <linux/mqueue.h>
#include <linux/audit.h>
#include <linux/personality.h>
#include <linux/time.h>
#include <linux/netlink.h>
#include <linux/compiler.h>
#include <asm/unistd.h>
#include <linux/security.h>
#include <linux/list.h>
#include <linux/binfmts.h>
#include <linux/highmem.h>
#include <linux/syscalls.h>
#include <asm/syscall.h>
#include <linux/capability.h>
#include <linux/fs_struct.h>
#include <linux/compat.h>
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
#include <linux/ctype.h>
#include <linux/string.h>
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
#include <linux/uaccess.h>
fsnotify: Move mark list head from object into dedicated structure Currently notification marks are attached to object (inode or vfsmnt) by a hlist_head in the object. The list is also protected by a spinlock in the object. So while there is any mark attached to the list of marks, the object must be pinned in memory (and thus e.g. last iput() deleting inode cannot happen). Also for list iteration in fsnotify() to work, we must hold fsnotify_mark_srcu lock so that mark itself and mark->obj_list.next cannot get freed. Thus we are required to wait for response to fanotify events from userspace process with fsnotify_mark_srcu lock held. That causes issues when userspace process is buggy and does not reply to some event - basically the whole notification subsystem gets eventually stuck. So to be able to drop fsnotify_mark_srcu lock while waiting for response, we have to pin the mark in memory and make sure it stays in the object list (as removing the mark waiting for response could lead to lost notification events for groups later in the list). However we don't want inode reclaim to block on such mark as that would lead to system just locking up elsewhere. This commit is the first in the series that paves way towards solving these conflicting lifetime needs. Instead of anchoring the list of marks directly in the object, we anchor it in a dedicated structure (fsnotify_mark_connector) and just point to that structure from the object. The following commits will also add spinlock protecting the list and object pointer to the structure. Reviewed-by: Miklos Szeredi <mszeredi@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-03-14 11:31:02 +00:00
#include <linux/fsnotify_backend.h>
#include <uapi/linux/limits.h>
audit: log nftables configuration change events iptables, ip6tables, arptables and ebtables table registration, replacement and unregistration configuration events are logged for the native (legacy) iptables setsockopt api, but not for the nftables netlink api which is used by the nft-variant of iptables in addition to nftables itself. Add calls to log the configuration actions in the nftables netlink api. This uses the same NETFILTER_CFG record format but overloads the table field. type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.878:162) : table=?:0;?:0 family=unspecified entries=2 op=nft_register_gen pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.878:162) : table=firewalld:1;?:0 family=inet entries=0 op=nft_register_table pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;filter_FORWARD:85 family=inet entries=8 op=nft_register_chain pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;filter_FORWARD:85 family=inet entries=101 op=nft_register_rule pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;__set0:87 family=inet entries=87 op=nft_register_setelem pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;__set0:87 family=inet entries=0 op=nft_register_set pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld For further information please see issue https://github.com/linux-audit/audit-kernel/issues/124 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-06-04 13:20:49 +00:00
#include <uapi/linux/netfilter/nf_tables.h>
audit: add OPENAT2 record to list "how" info Since the openat2(2) syscall uses a struct open_how pointer to communicate its parameters they are not usefully recorded by the audit SYSCALL record's four existing arguments. Add a new audit record type OPENAT2 that reports the parameters in its third argument, struct open_how with fields oflag, mode and resolve. The new record in the context of an event would look like: time->Wed Mar 17 16:28:53 2021 type=PROCTITLE msg=audit(1616012933.531:184): proctitle= 73797363616C6C735F66696C652F6F70656E617432002F746D702F61756469742D 7465737473756974652D737641440066696C652D6F70656E617432 type=PATH msg=audit(1616012933.531:184): item=1 name="file-openat2" inode=29 dev=00:1f mode=0100600 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmp_t:s0 nametype=CREATE cap_fp=0 cap_fi=0 cap_fe=0 cap_fver=0 cap_frootid=0 type=PATH msg=audit(1616012933.531:184): item=0 name="/root/rgb/git/audit-testsuite/tests" inode=25 dev=00:1f mode=040700 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmp_t:s0 nametype=PARENT cap_fp=0 cap_fi=0 cap_fe=0 cap_fver=0 cap_frootid=0 type=CWD msg=audit(1616012933.531:184): cwd="/root/rgb/git/audit-testsuite/tests" type=OPENAT2 msg=audit(1616012933.531:184): oflag=0100302 mode=0600 resolve=0xa type=SYSCALL msg=audit(1616012933.531:184): arch=c000003e syscall=437 success=yes exit=4 a0=3 a1=7ffe315f1c53 a2=7ffe315f1550 a3=18 items=2 ppid=528 pid=540 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=ttyS0 ses=1 comm="openat2" exe="/root/rgb/git/audit-testsuite/tests/syscalls_file/openat2" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key="testsuite-1616012933-bjAUcEPO" Link: https://lore.kernel.org/r/d23fbb89186754487850367224b060e26f9b7181.1621363275.git.rgb@redhat.com Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Christian Brauner <christian.brauner@ubuntu.com> [PM: tweak subject, wrap example, move AUDIT_OPENAT2 to 1337] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-05-19 20:00:22 +00:00
#include <uapi/linux/openat2.h> // struct open_how
#include <uapi/linux/fanotify.h>
#include "audit.h"
Audit: push audit success and retcode into arch ptrace.h The audit system previously expected arches calling to audit_syscall_exit to supply as arguments if the syscall was a success and what the return code was. Audit also provides a helper AUDITSC_RESULT which was supposed to simplify things by converting from negative retcodes to an audit internal magic value stating success or failure. This helper was wrong and could indicate that a valid pointer returned to userspace was a failed syscall. The fix is to fix the layering foolishness. We now pass audit_syscall_exit a struct pt_reg and it in turns calls back into arch code to collect the return value and to determine if the syscall was a success or failure. We also define a generic is_syscall_success() macro which determines success/failure based on if the value is < -MAX_ERRNO. This works for arches like x86 which do not use a separate mechanism to indicate syscall failure. We make both the is_syscall_success() and regs_return_value() static inlines instead of macros. The reason is because the audit function must take a void* for the regs. (uml calls theirs struct uml_pt_regs instead of just struct pt_regs so audit_syscall_exit can't take a struct pt_regs). Since the audit function takes a void* we need to use static inlines to cast it back to the arch correct structure to dereference it. The other major change is that on some arches, like ia64, MIPS and ppc, we change regs_return_value() to give us the negative value on syscall failure. THE only other user of this macro, kretprobe_example.c, won't notice and it makes the value signed consistently for the audit functions across all archs. In arch/sh/kernel/ptrace_64.c I see that we were using regs[9] in the old audit code as the return value. But the ptrace_64.h code defined the macro regs_return_value() as regs[3]. I have no idea which one is correct, but this patch now uses the regs_return_value() function, so it now uses regs[3]. For powerpc we previously used regs->result but now use the regs_return_value() function which uses regs->gprs[3]. regs->gprs[3] is always positive so the regs_return_value(), much like ia64 makes it negative before calling the audit code when appropriate. Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: H. Peter Anvin <hpa@zytor.com> [for x86 portion] Acked-by: Tony Luck <tony.luck@intel.com> [for ia64] Acked-by: Richard Weinberger <richard@nod.at> [for uml] Acked-by: David S. Miller <davem@davemloft.net> [for sparc] Acked-by: Ralf Baechle <ralf@linux-mips.org> [for mips] Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> [for ppc]
2012-01-03 19:23:06 +00:00
/* flags stating the success for a syscall */
#define AUDITSC_INVALID 0
#define AUDITSC_SUCCESS 1
#define AUDITSC_FAILURE 2
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* no execve audit message should be longer than this (userspace limits),
* see the note near the top of audit_log_execve_info() about this value */
#define MAX_EXECVE_AUDIT_LEN 7500
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
/* max length to print of cmdline/proctitle value during audit */
#define MAX_PROCTITLE_AUDIT_LEN 128
/* number of audit rules */
int audit_n_rules;
/* determines whether we collect data for signals sent */
int audit_signals;
struct audit_aux_data {
struct audit_aux_data *next;
int type;
};
/* Number of target pids per aux struct. */
#define AUDIT_AUX_PIDS 16
struct audit_aux_data_pids {
struct audit_aux_data d;
pid_t target_pid[AUDIT_AUX_PIDS];
kuid_t target_auid[AUDIT_AUX_PIDS];
kuid_t target_uid[AUDIT_AUX_PIDS];
unsigned int target_sessionid[AUDIT_AUX_PIDS];
u32 target_sid[AUDIT_AUX_PIDS];
char target_comm[AUDIT_AUX_PIDS][TASK_COMM_LEN];
int pid_count;
};
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
struct audit_aux_data_bprm_fcaps {
struct audit_aux_data d;
struct audit_cap_data fcap;
unsigned int fcap_ver;
struct audit_cap_data old_pcap;
struct audit_cap_data new_pcap;
};
struct audit_tree_refs {
struct audit_tree_refs *next;
struct audit_chunk *c[31];
};
struct audit_nfcfgop_tab {
enum audit_nfcfgop op;
const char *s;
};
static const struct audit_nfcfgop_tab audit_nfcfgs[] = {
audit: log nftables configuration change events iptables, ip6tables, arptables and ebtables table registration, replacement and unregistration configuration events are logged for the native (legacy) iptables setsockopt api, but not for the nftables netlink api which is used by the nft-variant of iptables in addition to nftables itself. Add calls to log the configuration actions in the nftables netlink api. This uses the same NETFILTER_CFG record format but overloads the table field. type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.878:162) : table=?:0;?:0 family=unspecified entries=2 op=nft_register_gen pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.878:162) : table=firewalld:1;?:0 family=inet entries=0 op=nft_register_table pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;filter_FORWARD:85 family=inet entries=8 op=nft_register_chain pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;filter_FORWARD:85 family=inet entries=101 op=nft_register_rule pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;__set0:87 family=inet entries=87 op=nft_register_setelem pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;__set0:87 family=inet entries=0 op=nft_register_set pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld For further information please see issue https://github.com/linux-audit/audit-kernel/issues/124 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-06-04 13:20:49 +00:00
{ AUDIT_XT_OP_REGISTER, "xt_register" },
{ AUDIT_XT_OP_REPLACE, "xt_replace" },
{ AUDIT_XT_OP_UNREGISTER, "xt_unregister" },
{ AUDIT_NFT_OP_TABLE_REGISTER, "nft_register_table" },
{ AUDIT_NFT_OP_TABLE_UNREGISTER, "nft_unregister_table" },
{ AUDIT_NFT_OP_CHAIN_REGISTER, "nft_register_chain" },
{ AUDIT_NFT_OP_CHAIN_UNREGISTER, "nft_unregister_chain" },
{ AUDIT_NFT_OP_RULE_REGISTER, "nft_register_rule" },
{ AUDIT_NFT_OP_RULE_UNREGISTER, "nft_unregister_rule" },
{ AUDIT_NFT_OP_SET_REGISTER, "nft_register_set" },
{ AUDIT_NFT_OP_SET_UNREGISTER, "nft_unregister_set" },
{ AUDIT_NFT_OP_SETELEM_REGISTER, "nft_register_setelem" },
{ AUDIT_NFT_OP_SETELEM_UNREGISTER, "nft_unregister_setelem" },
{ AUDIT_NFT_OP_GEN_REGISTER, "nft_register_gen" },
{ AUDIT_NFT_OP_OBJ_REGISTER, "nft_register_obj" },
{ AUDIT_NFT_OP_OBJ_UNREGISTER, "nft_unregister_obj" },
{ AUDIT_NFT_OP_OBJ_RESET, "nft_reset_obj" },
{ AUDIT_NFT_OP_FLOWTABLE_REGISTER, "nft_register_flowtable" },
{ AUDIT_NFT_OP_FLOWTABLE_UNREGISTER, "nft_unregister_flowtable" },
{ AUDIT_NFT_OP_SETELEM_RESET, "nft_reset_setelem" },
{ AUDIT_NFT_OP_RULE_RESET, "nft_reset_rule" },
audit: log nftables configuration change events iptables, ip6tables, arptables and ebtables table registration, replacement and unregistration configuration events are logged for the native (legacy) iptables setsockopt api, but not for the nftables netlink api which is used by the nft-variant of iptables in addition to nftables itself. Add calls to log the configuration actions in the nftables netlink api. This uses the same NETFILTER_CFG record format but overloads the table field. type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.878:162) : table=?:0;?:0 family=unspecified entries=2 op=nft_register_gen pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.878:162) : table=firewalld:1;?:0 family=inet entries=0 op=nft_register_table pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;filter_FORWARD:85 family=inet entries=8 op=nft_register_chain pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;filter_FORWARD:85 family=inet entries=101 op=nft_register_rule pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;__set0:87 family=inet entries=87 op=nft_register_setelem pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld ... type=NETFILTER_CFG msg=audit(2020-05-28 17:46:41.911:163) : table=firewalld:1;__set0:87 family=inet entries=0 op=nft_register_set pid=396 subj=system_u:system_r:firewalld_t:s0 comm=firewalld For further information please see issue https://github.com/linux-audit/audit-kernel/issues/124 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-06-04 13:20:49 +00:00
{ AUDIT_NFT_OP_INVALID, "nft_invalid" },
};
static int audit_match_perm(struct audit_context *ctx, int mask)
{
unsigned n;
if (unlikely(!ctx))
return 0;
n = ctx->major;
switch (audit_classify_syscall(ctx->arch, n)) {
case AUDITSC_NATIVE:
if ((mask & AUDIT_PERM_WRITE) &&
audit_match_class(AUDIT_CLASS_WRITE, n))
return 1;
if ((mask & AUDIT_PERM_READ) &&
audit_match_class(AUDIT_CLASS_READ, n))
return 1;
if ((mask & AUDIT_PERM_ATTR) &&
audit_match_class(AUDIT_CLASS_CHATTR, n))
return 1;
return 0;
case AUDITSC_COMPAT: /* 32bit on biarch */
if ((mask & AUDIT_PERM_WRITE) &&
audit_match_class(AUDIT_CLASS_WRITE_32, n))
return 1;
if ((mask & AUDIT_PERM_READ) &&
audit_match_class(AUDIT_CLASS_READ_32, n))
return 1;
if ((mask & AUDIT_PERM_ATTR) &&
audit_match_class(AUDIT_CLASS_CHATTR_32, n))
return 1;
return 0;
case AUDITSC_OPEN:
return mask & ACC_MODE(ctx->argv[1]);
case AUDITSC_OPENAT:
return mask & ACC_MODE(ctx->argv[2]);
case AUDITSC_SOCKETCALL:
return ((mask & AUDIT_PERM_WRITE) && ctx->argv[0] == SYS_BIND);
case AUDITSC_EXECVE:
return mask & AUDIT_PERM_EXEC;
case AUDITSC_OPENAT2:
return mask & ACC_MODE((u32)ctx->openat2.flags);
default:
return 0;
}
}
static int audit_match_filetype(struct audit_context *ctx, int val)
{
struct audit_names *n;
umode_t mode = (umode_t)val;
if (unlikely(!ctx))
return 0;
list_for_each_entry(n, &ctx->names_list, list) {
if ((n->ino != AUDIT_INO_UNSET) &&
((n->mode & S_IFMT) == mode))
return 1;
}
return 0;
}
/*
* We keep a linked list of fixed-sized (31 pointer) arrays of audit_chunk *;
* ->first_trees points to its beginning, ->trees - to the current end of data.
* ->tree_count is the number of free entries in array pointed to by ->trees.
* Original condition is (NULL, NULL, 0); as soon as it grows we never revert to NULL,
* "empty" becomes (p, p, 31) afterwards. We don't shrink the list (and seriously,
* it's going to remain 1-element for almost any setup) until we free context itself.
* References in it _are_ dropped - at the same time we free/drop aux stuff.
*/
static void audit_set_auditable(struct audit_context *ctx)
{
if (!ctx->prio) {
ctx->prio = 1;
ctx->current_state = AUDIT_STATE_RECORD;
}
}
static int put_tree_ref(struct audit_context *ctx, struct audit_chunk *chunk)
{
struct audit_tree_refs *p = ctx->trees;
int left = ctx->tree_count;
if (likely(left)) {
p->c[--left] = chunk;
ctx->tree_count = left;
return 1;
}
if (!p)
return 0;
p = p->next;
if (p) {
p->c[30] = chunk;
ctx->trees = p;
ctx->tree_count = 30;
return 1;
}
return 0;
}
static int grow_tree_refs(struct audit_context *ctx)
{
struct audit_tree_refs *p = ctx->trees;
ctx->trees = kzalloc(sizeof(struct audit_tree_refs), GFP_KERNEL);
if (!ctx->trees) {
ctx->trees = p;
return 0;
}
if (p)
p->next = ctx->trees;
else
ctx->first_trees = ctx->trees;
ctx->tree_count = 31;
return 1;
}
static void unroll_tree_refs(struct audit_context *ctx,
struct audit_tree_refs *p, int count)
{
struct audit_tree_refs *q;
int n;
if (!p) {
/* we started with empty chain */
p = ctx->first_trees;
count = 31;
/* if the very first allocation has failed, nothing to do */
if (!p)
return;
}
n = count;
for (q = p; q != ctx->trees; q = q->next, n = 31) {
while (n--) {
audit_put_chunk(q->c[n]);
q->c[n] = NULL;
}
}
while (n-- > ctx->tree_count) {
audit_put_chunk(q->c[n]);
q->c[n] = NULL;
}
ctx->trees = p;
ctx->tree_count = count;
}
static void free_tree_refs(struct audit_context *ctx)
{
struct audit_tree_refs *p, *q;
for (p = ctx->first_trees; p; p = q) {
q = p->next;
kfree(p);
}
}
static int match_tree_refs(struct audit_context *ctx, struct audit_tree *tree)
{
struct audit_tree_refs *p;
int n;
if (!tree)
return 0;
/* full ones */
for (p = ctx->first_trees; p != ctx->trees; p = p->next) {
for (n = 0; n < 31; n++)
if (audit_tree_match(p->c[n], tree))
return 1;
}
/* partial */
if (p) {
for (n = ctx->tree_count; n < 31; n++)
if (audit_tree_match(p->c[n], tree))
return 1;
}
return 0;
}
static int audit_compare_uid(kuid_t uid,
struct audit_names *name,
struct audit_field *f,
struct audit_context *ctx)
{
struct audit_names *n;
int rc;
if (name) {
rc = audit_uid_comparator(uid, f->op, name->uid);
if (rc)
return rc;
}
if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
rc = audit_uid_comparator(uid, f->op, n->uid);
if (rc)
return rc;
}
}
return 0;
}
static int audit_compare_gid(kgid_t gid,
struct audit_names *name,
struct audit_field *f,
struct audit_context *ctx)
{
struct audit_names *n;
int rc;
if (name) {
rc = audit_gid_comparator(gid, f->op, name->gid);
if (rc)
return rc;
}
if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
rc = audit_gid_comparator(gid, f->op, n->gid);
if (rc)
return rc;
}
}
return 0;
}
static int audit_field_compare(struct task_struct *tsk,
const struct cred *cred,
struct audit_field *f,
struct audit_context *ctx,
struct audit_names *name)
{
switch (f->val) {
/* process to file object comparisons */
case AUDIT_COMPARE_UID_TO_OBJ_UID:
return audit_compare_uid(cred->uid, name, f, ctx);
case AUDIT_COMPARE_GID_TO_OBJ_GID:
return audit_compare_gid(cred->gid, name, f, ctx);
case AUDIT_COMPARE_EUID_TO_OBJ_UID:
return audit_compare_uid(cred->euid, name, f, ctx);
case AUDIT_COMPARE_EGID_TO_OBJ_GID:
return audit_compare_gid(cred->egid, name, f, ctx);
case AUDIT_COMPARE_AUID_TO_OBJ_UID:
return audit_compare_uid(audit_get_loginuid(tsk), name, f, ctx);
case AUDIT_COMPARE_SUID_TO_OBJ_UID:
return audit_compare_uid(cred->suid, name, f, ctx);
case AUDIT_COMPARE_SGID_TO_OBJ_GID:
return audit_compare_gid(cred->sgid, name, f, ctx);
case AUDIT_COMPARE_FSUID_TO_OBJ_UID:
return audit_compare_uid(cred->fsuid, name, f, ctx);
case AUDIT_COMPARE_FSGID_TO_OBJ_GID:
return audit_compare_gid(cred->fsgid, name, f, ctx);
/* uid comparisons */
case AUDIT_COMPARE_UID_TO_AUID:
return audit_uid_comparator(cred->uid, f->op,
audit_get_loginuid(tsk));
case AUDIT_COMPARE_UID_TO_EUID:
return audit_uid_comparator(cred->uid, f->op, cred->euid);
case AUDIT_COMPARE_UID_TO_SUID:
return audit_uid_comparator(cred->uid, f->op, cred->suid);
case AUDIT_COMPARE_UID_TO_FSUID:
return audit_uid_comparator(cred->uid, f->op, cred->fsuid);
/* auid comparisons */
case AUDIT_COMPARE_AUID_TO_EUID:
return audit_uid_comparator(audit_get_loginuid(tsk), f->op,
cred->euid);
case AUDIT_COMPARE_AUID_TO_SUID:
return audit_uid_comparator(audit_get_loginuid(tsk), f->op,
cred->suid);
case AUDIT_COMPARE_AUID_TO_FSUID:
return audit_uid_comparator(audit_get_loginuid(tsk), f->op,
cred->fsuid);
/* euid comparisons */
case AUDIT_COMPARE_EUID_TO_SUID:
return audit_uid_comparator(cred->euid, f->op, cred->suid);
case AUDIT_COMPARE_EUID_TO_FSUID:
return audit_uid_comparator(cred->euid, f->op, cred->fsuid);
/* suid comparisons */
case AUDIT_COMPARE_SUID_TO_FSUID:
return audit_uid_comparator(cred->suid, f->op, cred->fsuid);
/* gid comparisons */
case AUDIT_COMPARE_GID_TO_EGID:
return audit_gid_comparator(cred->gid, f->op, cred->egid);
case AUDIT_COMPARE_GID_TO_SGID:
return audit_gid_comparator(cred->gid, f->op, cred->sgid);
case AUDIT_COMPARE_GID_TO_FSGID:
return audit_gid_comparator(cred->gid, f->op, cred->fsgid);
/* egid comparisons */
case AUDIT_COMPARE_EGID_TO_SGID:
return audit_gid_comparator(cred->egid, f->op, cred->sgid);
case AUDIT_COMPARE_EGID_TO_FSGID:
return audit_gid_comparator(cred->egid, f->op, cred->fsgid);
/* sgid comparison */
case AUDIT_COMPARE_SGID_TO_FSGID:
return audit_gid_comparator(cred->sgid, f->op, cred->fsgid);
default:
WARN(1, "Missing AUDIT_COMPARE define. Report as a bug\n");
return 0;
}
return 0;
}
/* Determine if any context name data matches a rule's watch data */
/* Compare a task_struct with an audit_rule. Return 1 on match, 0
* otherwise.
*
* If task_creation is true, this is an explicit indication that we are
* filtering a task rule at task creation time. This and tsk == current are
* the only situations where tsk->cred may be accessed without an rcu read lock.
*/
static int audit_filter_rules(struct task_struct *tsk,
struct audit_krule *rule,
struct audit_context *ctx,
struct audit_names *name,
enum audit_state *state,
bool task_creation)
{
const struct cred *cred;
int i, need_sid = 1;
u32 sid;
unsigned int sessionid;
if (ctx && rule->prio <= ctx->prio)
return 0;
cred = rcu_dereference_check(tsk->cred, tsk == current || task_creation);
for (i = 0; i < rule->field_count; i++) {
struct audit_field *f = &rule->fields[i];
struct audit_names *n;
int result = 0;
pid_t pid;
switch (f->type) {
case AUDIT_PID:
pid = task_tgid_nr(tsk);
result = audit_comparator(pid, f->op, f->val);
break;
case AUDIT_PPID:
if (ctx) {
if (!ctx->ppid)
ctx->ppid = task_ppid_nr(tsk);
result = audit_comparator(ctx->ppid, f->op, f->val);
}
break;
case AUDIT_EXE:
result = audit_exe_compare(tsk, rule->exe);
if (f->op == Audit_not_equal)
result = !result;
break;
case AUDIT_UID:
result = audit_uid_comparator(cred->uid, f->op, f->uid);
break;
case AUDIT_EUID:
result = audit_uid_comparator(cred->euid, f->op, f->uid);
break;
case AUDIT_SUID:
result = audit_uid_comparator(cred->suid, f->op, f->uid);
break;
case AUDIT_FSUID:
result = audit_uid_comparator(cred->fsuid, f->op, f->uid);
break;
case AUDIT_GID:
result = audit_gid_comparator(cred->gid, f->op, f->gid);
if (f->op == Audit_equal) {
if (!result)
result = groups_search(cred->group_info, f->gid);
} else if (f->op == Audit_not_equal) {
if (result)
result = !groups_search(cred->group_info, f->gid);
}
break;
case AUDIT_EGID:
result = audit_gid_comparator(cred->egid, f->op, f->gid);
if (f->op == Audit_equal) {
if (!result)
result = groups_search(cred->group_info, f->gid);
} else if (f->op == Audit_not_equal) {
if (result)
result = !groups_search(cred->group_info, f->gid);
}
break;
case AUDIT_SGID:
result = audit_gid_comparator(cred->sgid, f->op, f->gid);
break;
case AUDIT_FSGID:
result = audit_gid_comparator(cred->fsgid, f->op, f->gid);
break;
case AUDIT_SESSIONID:
sessionid = audit_get_sessionid(tsk);
result = audit_comparator(sessionid, f->op, f->val);
break;
case AUDIT_PERS:
result = audit_comparator(tsk->personality, f->op, f->val);
break;
case AUDIT_ARCH:
if (ctx)
result = audit_comparator(ctx->arch, f->op, f->val);
break;
case AUDIT_EXIT:
if (ctx && ctx->return_valid != AUDITSC_INVALID)
result = audit_comparator(ctx->return_code, f->op, f->val);
break;
case AUDIT_SUCCESS:
if (ctx && ctx->return_valid != AUDITSC_INVALID) {
if (f->val)
result = audit_comparator(ctx->return_valid, f->op, AUDITSC_SUCCESS);
else
result = audit_comparator(ctx->return_valid, f->op, AUDITSC_FAILURE);
}
break;
case AUDIT_DEVMAJOR:
if (name) {
if (audit_comparator(MAJOR(name->dev), f->op, f->val) ||
audit_comparator(MAJOR(name->rdev), f->op, f->val))
++result;
} else if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
if (audit_comparator(MAJOR(n->dev), f->op, f->val) ||
audit_comparator(MAJOR(n->rdev), f->op, f->val)) {
++result;
break;
}
}
}
break;
case AUDIT_DEVMINOR:
if (name) {
if (audit_comparator(MINOR(name->dev), f->op, f->val) ||
audit_comparator(MINOR(name->rdev), f->op, f->val))
++result;
} else if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
if (audit_comparator(MINOR(n->dev), f->op, f->val) ||
audit_comparator(MINOR(n->rdev), f->op, f->val)) {
++result;
break;
}
}
}
break;
case AUDIT_INODE:
if (name)
result = audit_comparator(name->ino, f->op, f->val);
else if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
if (audit_comparator(n->ino, f->op, f->val)) {
++result;
break;
}
}
}
break;
case AUDIT_OBJ_UID:
if (name) {
result = audit_uid_comparator(name->uid, f->op, f->uid);
} else if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
if (audit_uid_comparator(n->uid, f->op, f->uid)) {
++result;
break;
}
}
}
break;
case AUDIT_OBJ_GID:
if (name) {
result = audit_gid_comparator(name->gid, f->op, f->gid);
} else if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
if (audit_gid_comparator(n->gid, f->op, f->gid)) {
++result;
break;
}
}
}
break;
case AUDIT_WATCH:
if (name) {
result = audit_watch_compare(rule->watch,
name->ino,
name->dev);
if (f->op == Audit_not_equal)
result = !result;
}
break;
case AUDIT_DIR:
if (ctx) {
result = match_tree_refs(ctx, rule->tree);
if (f->op == Audit_not_equal)
result = !result;
}
break;
case AUDIT_LOGINUID:
result = audit_uid_comparator(audit_get_loginuid(tsk),
f->op, f->uid);
break;
case AUDIT_LOGINUID_SET:
result = audit_comparator(audit_loginuid_set(tsk), f->op, f->val);
break;
case AUDIT_SADDR_FAM:
if (ctx && ctx->sockaddr)
result = audit_comparator(ctx->sockaddr->ss_family,
f->op, f->val);
break;
case AUDIT_SUBJ_USER:
case AUDIT_SUBJ_ROLE:
case AUDIT_SUBJ_TYPE:
case AUDIT_SUBJ_SEN:
case AUDIT_SUBJ_CLR:
/* NOTE: this may return negative values indicating
a temporary error. We simply treat this as a
match for now to avoid losing information that
may be wanted. An error message will also be
logged upon error */
if (f->lsm_rule) {
if (need_sid) {
/* @tsk should always be equal to
* @current with the exception of
* fork()/copy_process() in which case
* the new @tsk creds are still a dup
* of @current's creds so we can still
* use security_current_getsecid_subj()
* here even though it always refs
* @current's creds
*/
security_current_getsecid_subj(&sid);
need_sid = 0;
}
result = security_audit_rule_match(sid, f->type,
f->op,
f->lsm_rule);
}
break;
case AUDIT_OBJ_USER:
case AUDIT_OBJ_ROLE:
case AUDIT_OBJ_TYPE:
case AUDIT_OBJ_LEV_LOW:
case AUDIT_OBJ_LEV_HIGH:
/* The above note for AUDIT_SUBJ_USER...AUDIT_SUBJ_CLR
also applies here */
if (f->lsm_rule) {
/* Find files that match */
if (name) {
result = security_audit_rule_match(
name->osid,
f->type,
f->op,
f->lsm_rule);
} else if (ctx) {
list_for_each_entry(n, &ctx->names_list, list) {
if (security_audit_rule_match(
n->osid,
f->type,
f->op,
f->lsm_rule)) {
++result;
break;
}
}
}
/* Find ipc objects that match */
if (!ctx || ctx->type != AUDIT_IPC)
break;
if (security_audit_rule_match(ctx->ipc.osid,
f->type, f->op,
f->lsm_rule))
++result;
}
break;
case AUDIT_ARG0:
case AUDIT_ARG1:
case AUDIT_ARG2:
case AUDIT_ARG3:
if (ctx)
result = audit_comparator(ctx->argv[f->type-AUDIT_ARG0], f->op, f->val);
break;
case AUDIT_FILTERKEY:
/* ignore this field for filtering */
result = 1;
break;
case AUDIT_PERM:
result = audit_match_perm(ctx, f->val);
if (f->op == Audit_not_equal)
result = !result;
break;
case AUDIT_FILETYPE:
result = audit_match_filetype(ctx, f->val);
if (f->op == Audit_not_equal)
result = !result;
break;
case AUDIT_FIELD_COMPARE:
result = audit_field_compare(tsk, cred, f, ctx, name);
break;
}
if (!result)
return 0;
}
if (ctx) {
if (rule->filterkey) {
kfree(ctx->filterkey);
ctx->filterkey = kstrdup(rule->filterkey, GFP_ATOMIC);
}
ctx->prio = rule->prio;
}
switch (rule->action) {
case AUDIT_NEVER:
*state = AUDIT_STATE_DISABLED;
break;
case AUDIT_ALWAYS:
*state = AUDIT_STATE_RECORD;
break;
}
return 1;
}
/* At process creation time, we can determine if system-call auditing is
* completely disabled for this task. Since we only have the task
* structure at this point, we can only check uid and gid.
*/
static enum audit_state audit_filter_task(struct task_struct *tsk, char **key)
{
struct audit_entry *e;
enum audit_state state;
rcu_read_lock();
list_for_each_entry_rcu(e, &audit_filter_list[AUDIT_FILTER_TASK], list) {
if (audit_filter_rules(tsk, &e->rule, NULL, NULL,
&state, true)) {
if (state == AUDIT_STATE_RECORD)
*key = kstrdup(e->rule.filterkey, GFP_ATOMIC);
rcu_read_unlock();
return state;
}
}
rcu_read_unlock();
return AUDIT_STATE_BUILD;
}
static int audit_in_mask(const struct audit_krule *rule, unsigned long val)
{
int word, bit;
if (val > 0xffffffff)
return false;
word = AUDIT_WORD(val);
if (word >= AUDIT_BITMASK_SIZE)
return false;
bit = AUDIT_BIT(val);
return rule->mask[word] & bit;
}
/**
* __audit_filter_op - common filter helper for operations (syscall/uring/etc)
* @tsk: associated task
* @ctx: audit context
* @list: audit filter list
* @name: audit_name (can be NULL)
* @op: current syscall/uring_op
*
* Run the udit filters specified in @list against @tsk using @ctx,
* @name, and @op, as necessary; the caller is responsible for ensuring
* that the call is made while the RCU read lock is held. The @name
* parameter can be NULL, but all others must be specified.
* Returns 1/true if the filter finds a match, 0/false if none are found.
*/
static int __audit_filter_op(struct task_struct *tsk,
struct audit_context *ctx,
struct list_head *list,
struct audit_names *name,
unsigned long op)
{
struct audit_entry *e;
enum audit_state state;
list_for_each_entry_rcu(e, list, list) {
if (audit_in_mask(&e->rule, op) &&
audit_filter_rules(tsk, &e->rule, ctx, name,
&state, false)) {
ctx->current_state = state;
return 1;
}
}
return 0;
}
/**
* audit_filter_uring - apply filters to an io_uring operation
* @tsk: associated task
* @ctx: audit context
*/
static void audit_filter_uring(struct task_struct *tsk,
struct audit_context *ctx)
{
if (auditd_test_task(tsk))
return;
rcu_read_lock();
__audit_filter_op(tsk, ctx, &audit_filter_list[AUDIT_FILTER_URING_EXIT],
NULL, ctx->uring_op);
rcu_read_unlock();
}
/* At syscall exit time, this filter is called if the audit_state is
* not low enough that auditing cannot take place, but is also not
* high enough that we already know we have to write an audit record
* (i.e., the state is AUDIT_STATE_BUILD).
*/
static void audit_filter_syscall(struct task_struct *tsk,
struct audit_context *ctx)
{
audit: fix auditd/kernel connection state tracking What started as a rather straightforward race condition reported by Dmitry using the syzkaller fuzzer ended up revealing some major problems with how the audit subsystem managed its netlink sockets and its connection with the userspace audit daemon. Fixing this properly had quite the cascading effect and what we are left with is this rather large and complicated patch. My initial goal was to try and decompose this patch into multiple smaller patches, but the way these changes are intertwined makes it difficult to split these changes into meaningful pieces that don't break or somehow make things worse for the intermediate states. The patch makes a number of changes, but the most significant are highlighted below: * The auditd tracking variables, e.g. audit_sock, are now gone and replaced by a RCU/spin_lock protected variable auditd_conn which is a structure containing all of the auditd tracking information. * We no longer track the auditd sock directly, instead we track it via the network namespace in which it resides and we use the audit socket associated with that namespace. In spirit, this is what the code was trying to do prior to this patch (at least I think that is what the original authors intended), but it was done rather poorly and added a layer of obfuscation that only masked the underlying problems. * Big backlog queue cleanup, again. In v4.10 we made some pretty big changes to how the audit backlog queues work, here we haven't changed the queue design so much as cleaned up the implementation. Brought about by the locking changes, we've simplified kauditd_thread() quite a bit by consolidating the queue handling into a new helper function, kauditd_send_queue(), which allows us to eliminate a lot of very similar code and makes the looping logic in kauditd_thread() clearer. * All netlink messages sent to auditd are now sent via auditd_send_unicast_skb(). Other than just making sense, this makes the lock handling easier. * Change the audit_log_start() sleep behavior so that we never sleep on auditd events (unchanged) or if the caller is holding the audit_cmd_mutex (changed). Previously we didn't sleep if the caller was auditd or if the message type fell between a certain range; the type check was a poor effort of doing what the cmd_mutex check now does. Richard Guy Briggs originally proposed not sleeping the cmd_mutex owner several years ago but his patch wasn't acceptable at the time. At least the idea lives on here. * A problem with the lost record counter has been resolved. Steve Grubb and I both happened to notice this problem and according to some quick testing by Steve, this problem goes back quite some time. It's largely a harmless problem, although it may have left some careful sysadmins quite puzzled. Cc: <stable@vger.kernel.org> # 4.10.x- Reported-by: Dmitry Vyukov <dvyukov@google.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-03-21 15:26:35 +00:00
if (auditd_test_task(tsk))
return;
rcu_read_lock();
__audit_filter_op(tsk, ctx, &audit_filter_list[AUDIT_FILTER_EXIT],
NULL, ctx->major);
rcu_read_unlock();
}
/*
* Given an audit_name check the inode hash table to see if they match.
* Called holding the rcu read lock to protect the use of audit_inode_hash
*/
static int audit_filter_inode_name(struct task_struct *tsk,
struct audit_names *n,
struct audit_context *ctx)
{
int h = audit_hash_ino((u32)n->ino);
struct list_head *list = &audit_inode_hash[h];
return __audit_filter_op(tsk, ctx, list, n, ctx->major);
}
/* At syscall exit time, this filter is called if any audit_names have been
* collected during syscall processing. We only check rules in sublists at hash
* buckets applicable to the inode numbers in audit_names.
* Regarding audit_state, same rules apply as for audit_filter_syscall().
*/
void audit_filter_inodes(struct task_struct *tsk, struct audit_context *ctx)
{
struct audit_names *n;
audit: fix auditd/kernel connection state tracking What started as a rather straightforward race condition reported by Dmitry using the syzkaller fuzzer ended up revealing some major problems with how the audit subsystem managed its netlink sockets and its connection with the userspace audit daemon. Fixing this properly had quite the cascading effect and what we are left with is this rather large and complicated patch. My initial goal was to try and decompose this patch into multiple smaller patches, but the way these changes are intertwined makes it difficult to split these changes into meaningful pieces that don't break or somehow make things worse for the intermediate states. The patch makes a number of changes, but the most significant are highlighted below: * The auditd tracking variables, e.g. audit_sock, are now gone and replaced by a RCU/spin_lock protected variable auditd_conn which is a structure containing all of the auditd tracking information. * We no longer track the auditd sock directly, instead we track it via the network namespace in which it resides and we use the audit socket associated with that namespace. In spirit, this is what the code was trying to do prior to this patch (at least I think that is what the original authors intended), but it was done rather poorly and added a layer of obfuscation that only masked the underlying problems. * Big backlog queue cleanup, again. In v4.10 we made some pretty big changes to how the audit backlog queues work, here we haven't changed the queue design so much as cleaned up the implementation. Brought about by the locking changes, we've simplified kauditd_thread() quite a bit by consolidating the queue handling into a new helper function, kauditd_send_queue(), which allows us to eliminate a lot of very similar code and makes the looping logic in kauditd_thread() clearer. * All netlink messages sent to auditd are now sent via auditd_send_unicast_skb(). Other than just making sense, this makes the lock handling easier. * Change the audit_log_start() sleep behavior so that we never sleep on auditd events (unchanged) or if the caller is holding the audit_cmd_mutex (changed). Previously we didn't sleep if the caller was auditd or if the message type fell between a certain range; the type check was a poor effort of doing what the cmd_mutex check now does. Richard Guy Briggs originally proposed not sleeping the cmd_mutex owner several years ago but his patch wasn't acceptable at the time. At least the idea lives on here. * A problem with the lost record counter has been resolved. Steve Grubb and I both happened to notice this problem and according to some quick testing by Steve, this problem goes back quite some time. It's largely a harmless problem, although it may have left some careful sysadmins quite puzzled. Cc: <stable@vger.kernel.org> # 4.10.x- Reported-by: Dmitry Vyukov <dvyukov@google.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-03-21 15:26:35 +00:00
if (auditd_test_task(tsk))
return;
rcu_read_lock();
list_for_each_entry(n, &ctx->names_list, list) {
if (audit_filter_inode_name(tsk, n, ctx))
break;
}
rcu_read_unlock();
}
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
static inline void audit_proctitle_free(struct audit_context *context)
{
kfree(context->proctitle.value);
context->proctitle.value = NULL;
context->proctitle.len = 0;
}
static inline void audit_free_module(struct audit_context *context)
{
if (context->type == AUDIT_KERN_MODULE) {
kfree(context->module.name);
context->module.name = NULL;
}
}
static inline void audit_free_names(struct audit_context *context)
{
struct audit_names *n, *next;
list_for_each_entry_safe(n, next, &context->names_list, list) {
list_del(&n->list);
if (n->name)
putname(n->name);
if (n->should_free)
kfree(n);
}
context->name_count = 0;
path_put(&context->pwd);
context->pwd.dentry = NULL;
context->pwd.mnt = NULL;
}
static inline void audit_free_aux(struct audit_context *context)
{
struct audit_aux_data *aux;
while ((aux = context->aux)) {
context->aux = aux->next;
kfree(aux);
}
context->aux = NULL;
while ((aux = context->aux_pids)) {
context->aux_pids = aux->next;
kfree(aux);
}
context->aux_pids = NULL;
}
/**
* audit_reset_context - reset a audit_context structure
* @ctx: the audit_context to reset
*
* All fields in the audit_context will be reset to an initial state, all
* references held by fields will be dropped, and private memory will be
* released. When this function returns the audit_context will be suitable
* for reuse, so long as the passed context is not NULL or a dummy context.
*/
static void audit_reset_context(struct audit_context *ctx)
{
if (!ctx)
return;
/* if ctx is non-null, reset the "ctx->context" regardless */
ctx->context = AUDIT_CTX_UNUSED;
if (ctx->dummy)
return;
/*
* NOTE: It shouldn't matter in what order we release the fields, so
* release them in the order in which they appear in the struct;
* this gives us some hope of quickly making sure we are
* resetting the audit_context properly.
*
* Other things worth mentioning:
* - we don't reset "dummy"
* - we don't reset "state", we do reset "current_state"
* - we preserve "filterkey" if "state" is AUDIT_STATE_RECORD
* - much of this is likely overkill, but play it safe for now
* - we really need to work on improving the audit_context struct
*/
ctx->current_state = ctx->state;
ctx->serial = 0;
ctx->major = 0;
ctx->uring_op = 0;
ctx->ctime = (struct timespec64){ .tv_sec = 0, .tv_nsec = 0 };
memset(ctx->argv, 0, sizeof(ctx->argv));
ctx->return_code = 0;
ctx->prio = (ctx->state == AUDIT_STATE_RECORD ? ~0ULL : 0);
ctx->return_valid = AUDITSC_INVALID;
audit_free_names(ctx);
if (ctx->state != AUDIT_STATE_RECORD) {
kfree(ctx->filterkey);
ctx->filterkey = NULL;
}
audit_free_aux(ctx);
kfree(ctx->sockaddr);
ctx->sockaddr = NULL;
ctx->sockaddr_len = 0;
ctx->ppid = 0;
ctx->uid = ctx->euid = ctx->suid = ctx->fsuid = KUIDT_INIT(0);
ctx->gid = ctx->egid = ctx->sgid = ctx->fsgid = KGIDT_INIT(0);
ctx->personality = 0;
ctx->arch = 0;
ctx->target_pid = 0;
ctx->target_auid = ctx->target_uid = KUIDT_INIT(0);
ctx->target_sessionid = 0;
ctx->target_sid = 0;
ctx->target_comm[0] = '\0';
unroll_tree_refs(ctx, NULL, 0);
WARN_ON(!list_empty(&ctx->killed_trees));
audit_free_module(ctx);
ctx->fds[0] = -1;
ctx->type = 0; /* reset last for audit_free_*() */
}
static inline struct audit_context *audit_alloc_context(enum audit_state state)
{
struct audit_context *context;
context = kzalloc(sizeof(*context), GFP_KERNEL);
if (!context)
return NULL;
context->context = AUDIT_CTX_UNUSED;
context->state = state;
context->prio = state == AUDIT_STATE_RECORD ? ~0ULL : 0;
INIT_LIST_HEAD(&context->killed_trees);
INIT_LIST_HEAD(&context->names_list);
audit: trigger accompanying records when no rules present When there are no audit rules registered, mandatory records (config, etc.) are missing their accompanying records (syscall, proctitle, etc.). This is due to audit context dummy set on syscall entry based on absence of rules that signals that no other records are to be printed. Clear the dummy bit if any record is generated, open coding this in audit_log_start(). The proctitle context and dummy checks are pointless since the proctitle record will not be printed if no syscall records are printed. The fds array is reset to -1 after the first syscall to indicate it isn't valid any more, but was never set to -1 when the context was allocated to indicate it wasn't yet valid. Check ctx->pwd in audit_log_name(). The audit_inode* functions can be called without going through getname_flags() or getname_kernel() that sets audit_names and cwd, so set the cwd in audit_alloc_name() if it has not already been done so due to audit_names being valid and purge all other audit_getcwd() calls. Revert the LSM dump_common_audit_data() LSM_AUDIT_DATA_* cases from the ghak96 patch since they are no longer necessary due to cwd coverage in audit_alloc_name(). Thanks to bauen1 <j2468h@googlemail.com> for reporting LSM situations in which context->cwd is not valid, inadvertantly fixed by the ghak96 patch. Please see upstream github issue https://github.com/linux-audit/audit-kernel/issues/120 This is also related to upstream github issue https://github.com/linux-audit/audit-kernel/issues/96 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-09-22 12:44:50 +00:00
context->fds[0] = -1;
context->return_valid = AUDITSC_INVALID;
return context;
}
/**
* audit_alloc - allocate an audit context block for a task
* @tsk: task
*
* Filter on the task information and allocate a per-task audit context
* if necessary. Doing so turns on system call auditing for the
* specified task. This is called from copy_process, so no lock is
* needed.
*/
int audit_alloc(struct task_struct *tsk)
{
struct audit_context *context;
enum audit_state state;
char *key = NULL;
if (likely(!audit_ever_enabled))
return 0;
state = audit_filter_task(tsk, &key);
if (state == AUDIT_STATE_DISABLED) {
clear_task_syscall_work(tsk, SYSCALL_AUDIT);
return 0;
}
context = audit_alloc_context(state);
if (!context) {
kfree(key);
audit_log_lost("out of memory in audit_alloc");
return -ENOMEM;
}
context->filterkey = key;
audit_set_context(tsk, context);
set_task_syscall_work(tsk, SYSCALL_AUDIT);
return 0;
}
static inline void audit_free_context(struct audit_context *context)
{
/* resetting is extra work, but it is likely just noise */
audit_reset_context(context);
audit_proctitle_free(context);
free_tree_refs(context);
kfree(context->filterkey);
kfree(context);
}
static int audit_log_pid_context(struct audit_context *context, pid_t pid,
kuid_t auid, kuid_t uid, unsigned int sessionid,
u32 sid, char *comm)
{
struct audit_buffer *ab;
char *ctx = NULL;
u32 len;
int rc = 0;
ab = audit_log_start(context, GFP_KERNEL, AUDIT_OBJ_PID);
if (!ab)
return rc;
audit_log_format(ab, "opid=%d oauid=%d ouid=%d oses=%d", pid,
from_kuid(&init_user_ns, auid),
from_kuid(&init_user_ns, uid), sessionid);
if (sid) {
if (security_secid_to_secctx(sid, &ctx, &len)) {
audit_log_format(ab, " obj=(none)");
rc = 1;
} else {
audit_log_format(ab, " obj=%s", ctx);
security_release_secctx(ctx, len);
}
}
audit_log_format(ab, " ocomm=");
audit_log_untrustedstring(ab, comm);
audit_log_end(ab);
return rc;
}
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
static void audit_log_execve_info(struct audit_context *context,
struct audit_buffer **ab)
{
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
long len_max;
long len_rem;
long len_full;
long len_buf;
long len_abuf = 0;
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
long len_tmp;
bool require_data;
bool encode;
unsigned int iter;
unsigned int arg;
char *buf_head;
char *buf;
const char __user *p = (const char __user *)current->mm->arg_start;
/* NOTE: this buffer needs to be large enough to hold all the non-arg
* data we put in the audit record for this argument (see the
* code below) ... at this point in time 96 is plenty */
char abuf[96];
/* NOTE: we set MAX_EXECVE_AUDIT_LEN to a rather arbitrary limit, the
* current value of 7500 is not as important as the fact that it
* is less than 8k, a setting of 7500 gives us plenty of wiggle
* room if we go over a little bit in the logging below */
WARN_ON_ONCE(MAX_EXECVE_AUDIT_LEN > 7500);
len_max = MAX_EXECVE_AUDIT_LEN;
/* scratch buffer to hold the userspace args */
buf_head = kmalloc(MAX_EXECVE_AUDIT_LEN + 1, GFP_KERNEL);
if (!buf_head) {
audit_panic("out of memory for argv string");
return;
}
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
buf = buf_head;
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
audit_log_format(*ab, "argc=%d", context->execve.argc);
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
len_rem = len_max;
len_buf = 0;
len_full = 0;
require_data = true;
encode = false;
iter = 0;
arg = 0;
do {
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* NOTE: we don't ever want to trust this value for anything
* serious, but the audit record format insists we
* provide an argument length for really long arguments,
* e.g. > MAX_EXECVE_AUDIT_LEN, so we have no choice but
* to use strncpy_from_user() to obtain this value for
* recording in the log, although we don't use it
* anywhere here to avoid a double-fetch problem */
if (len_full == 0)
len_full = strnlen_user(p, MAX_ARG_STRLEN) - 1;
/* read more data from userspace */
if (require_data) {
/* can we make more room in the buffer? */
if (buf != buf_head) {
memmove(buf_head, buf, len_buf);
buf = buf_head;
}
/* fetch as much as we can of the argument */
len_tmp = strncpy_from_user(&buf_head[len_buf], p,
len_max - len_buf);
if (len_tmp == -EFAULT) {
/* unable to copy from userspace */
send_sig(SIGKILL, current, 0);
goto out;
} else if (len_tmp == (len_max - len_buf)) {
/* buffer is not large enough */
require_data = true;
/* NOTE: if we are going to span multiple
* buffers force the encoding so we stand
* a chance at a sane len_full value and
* consistent record encoding */
encode = true;
len_full = len_full * 2;
p += len_tmp;
} else {
require_data = false;
if (!encode)
encode = audit_string_contains_control(
buf, len_tmp);
/* try to use a trusted value for len_full */
if (len_full < len_max)
len_full = (encode ?
len_tmp * 2 : len_tmp);
p += len_tmp + 1;
}
len_buf += len_tmp;
buf_head[len_buf] = '\0';
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* length of the buffer in the audit record? */
len_abuf = (encode ? len_buf * 2 : len_buf + 2);
}
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* write as much as we can to the audit log */
if (len_buf >= 0) {
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* NOTE: some magic numbers here - basically if we
* can't fit a reasonable amount of data into the
* existing audit buffer, flush it and start with
* a new buffer */
if ((sizeof(abuf) + 8) > len_rem) {
len_rem = len_max;
audit_log_end(*ab);
*ab = audit_log_start(context,
GFP_KERNEL, AUDIT_EXECVE);
if (!*ab)
goto out;
}
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* create the non-arg portion of the arg record */
len_tmp = 0;
if (require_data || (iter > 0) ||
((len_abuf + sizeof(abuf)) > len_rem)) {
if (iter == 0) {
len_tmp += snprintf(&abuf[len_tmp],
sizeof(abuf) - len_tmp,
" a%d_len=%lu",
arg, len_full);
}
len_tmp += snprintf(&abuf[len_tmp],
sizeof(abuf) - len_tmp,
" a%d[%d]=", arg, iter++);
} else
len_tmp += snprintf(&abuf[len_tmp],
sizeof(abuf) - len_tmp,
" a%d=", arg);
WARN_ON(len_tmp >= sizeof(abuf));
abuf[sizeof(abuf) - 1] = '\0';
/* log the arg in the audit record */
audit_log_format(*ab, "%s", abuf);
len_rem -= len_tmp;
len_tmp = len_buf;
if (encode) {
if (len_abuf > len_rem)
len_tmp = len_rem / 2; /* encoding */
audit_log_n_hex(*ab, buf, len_tmp);
len_rem -= len_tmp * 2;
len_abuf -= len_tmp * 2;
} else {
if (len_abuf > len_rem)
len_tmp = len_rem - 2; /* quotes */
audit_log_n_string(*ab, buf, len_tmp);
len_rem -= len_tmp + 2;
/* don't subtract the "2" because we still need
* to add quotes to the remaining string */
len_abuf -= len_tmp;
}
len_buf -= len_tmp;
buf += len_tmp;
}
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* ready to move to the next argument? */
if ((len_buf == 0) && !require_data) {
arg++;
iter = 0;
len_full = 0;
require_data = true;
encode = false;
}
} while (arg < context->execve.argc);
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
/* NOTE: the caller handles the final audit_log_end() call */
audit: fix a double fetch in audit_log_single_execve_arg() There is a double fetch problem in audit_log_single_execve_arg() where we first check the execve(2) argumnets for any "bad" characters which would require hex encoding and then re-fetch the arguments for logging in the audit record[1]. Of course this leaves a window of opportunity for an unsavory application to munge with the data. This patch reworks things by only fetching the argument data once[2] into a buffer where it is scanned and logged into the audit records(s). In addition to fixing the double fetch, this patch improves on the original code in a few other ways: better handling of large arguments which require encoding, stricter record length checking, and some performance improvements (completely unverified, but we got rid of some strlen() calls, that's got to be a good thing). As part of the development of this patch, I've also created a basic regression test for the audit-testsuite, the test can be tracked on GitHub at the following link: * https://github.com/linux-audit/audit-testsuite/issues/25 [1] If you pay careful attention, there is actually a triple fetch problem due to a strnlen_user() call at the top of the function. [2] This is a tiny white lie, we do make a call to strnlen_user() prior to fetching the argument data. I don't like it, but due to the way the audit record is structured we really have no choice unless we copy the entire argument at once (which would require a rather wasteful allocation). The good news is that with this patch the kernel no longer relies on this strnlen_user() value for anything beyond recording it in the log, we also update it with a trustworthy value whenever possible. Reported-by: Pengfei Wang <wpengfeinudt@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2016-07-19 21:42:57 +00:00
out:
kfree(buf_head);
}
static void audit_log_cap(struct audit_buffer *ab, char *prefix,
kernel_cap_t *cap)
{
if (cap_isclear(*cap)) {
audit_log_format(ab, " %s=0", prefix);
return;
}
capability: just use a 'u64' instead of a 'u32[2]' array Back in 2008 we extended the capability bits from 32 to 64, and we did it by extending the single 32-bit capability word from one word to an array of two words. It was then obfuscated by hiding the "2" behind two macro expansions, with the reasoning being that maybe it gets extended further some day. That reasoning may have been valid at the time, but the last thing we want to do is to extend the capability set any more. And the array of values not only causes source code oddities (with loops to deal with it), but also results in worse code generation. It's a lose-lose situation. So just change the 'u32[2]' into a 'u64' and be done with it. We still have to deal with the fact that the user space interface is designed around an array of these 32-bit values, but that was the case before too, since the array layouts were different (ie user space doesn't use an array of 32-bit values for individual capability masks, but an array of 32-bit slices of multiple masks). So that marshalling of data is actually simplified too, even if it does remain somewhat obscure and odd. This was all triggered by my reaction to the new "cap_isidentical()" introduced recently. By just using a saner data structure, it went from unsigned __capi; CAP_FOR_EACH_U32(__capi) { if (a.cap[__capi] != b.cap[__capi]) return false; } return true; to just being return a.val == b.val; instead. Which is rather more obvious both to humans and to compilers. Cc: Mateusz Guzik <mjguzik@gmail.com> Cc: Casey Schaufler <casey@schaufler-ca.com> Cc: Serge Hallyn <serge@hallyn.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Paul Moore <paul@paul-moore.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2023-02-28 19:39:09 +00:00
audit_log_format(ab, " %s=%016llx", prefix, cap->val);
}
static void audit_log_fcaps(struct audit_buffer *ab, struct audit_names *name)
{
if (name->fcap_ver == -1) {
audit_log_format(ab, " cap_fe=? cap_fver=? cap_fp=? cap_fi=?");
return;
}
audit_log_cap(ab, "cap_fp", &name->fcap.permitted);
audit_log_cap(ab, "cap_fi", &name->fcap.inheritable);
audit_log_format(ab, " cap_fe=%d cap_fver=%x cap_frootid=%d",
name->fcap.fE, name->fcap_ver,
from_kuid(&init_user_ns, name->fcap.rootid));
}
static void audit_log_time(struct audit_context *context, struct audit_buffer **ab)
{
const struct audit_ntp_data *ntp = &context->time.ntp_data;
const struct timespec64 *tk = &context->time.tk_injoffset;
static const char * const ntp_name[] = {
"offset",
"freq",
"status",
"tai",
"tick",
"adjust",
};
int type;
if (context->type == AUDIT_TIME_ADJNTPVAL) {
for (type = 0; type < AUDIT_NTP_NVALS; type++) {
if (ntp->vals[type].newval != ntp->vals[type].oldval) {
if (!*ab) {
*ab = audit_log_start(context,
GFP_KERNEL,
AUDIT_TIME_ADJNTPVAL);
if (!*ab)
return;
}
audit_log_format(*ab, "op=%s old=%lli new=%lli",
ntp_name[type],
ntp->vals[type].oldval,
ntp->vals[type].newval);
audit_log_end(*ab);
*ab = NULL;
}
}
}
if (tk->tv_sec != 0 || tk->tv_nsec != 0) {
if (!*ab) {
*ab = audit_log_start(context, GFP_KERNEL,
AUDIT_TIME_INJOFFSET);
if (!*ab)
return;
}
audit_log_format(*ab, "sec=%lli nsec=%li",
(long long)tk->tv_sec, tk->tv_nsec);
audit_log_end(*ab);
*ab = NULL;
}
}
static void show_special(struct audit_context *context, int *call_panic)
{
struct audit_buffer *ab;
int i;
ab = audit_log_start(context, GFP_KERNEL, context->type);
if (!ab)
return;
switch (context->type) {
case AUDIT_SOCKETCALL: {
int nargs = context->socketcall.nargs;
audit_log_format(ab, "nargs=%d", nargs);
for (i = 0; i < nargs; i++)
audit_log_format(ab, " a%d=%lx", i,
context->socketcall.args[i]);
break; }
case AUDIT_IPC: {
u32 osid = context->ipc.osid;
audit_log_format(ab, "ouid=%u ogid=%u mode=%#ho",
from_kuid(&init_user_ns, context->ipc.uid),
from_kgid(&init_user_ns, context->ipc.gid),
context->ipc.mode);
if (osid) {
char *ctx = NULL;
u32 len;
if (security_secid_to_secctx(osid, &ctx, &len)) {
audit_log_format(ab, " osid=%u", osid);
*call_panic = 1;
} else {
audit_log_format(ab, " obj=%s", ctx);
security_release_secctx(ctx, len);
}
}
if (context->ipc.has_perm) {
audit_log_end(ab);
ab = audit_log_start(context, GFP_KERNEL,
AUDIT_IPC_SET_PERM);
if (unlikely(!ab))
return;
audit_log_format(ab,
"qbytes=%lx ouid=%u ogid=%u mode=%#ho",
context->ipc.qbytes,
context->ipc.perm_uid,
context->ipc.perm_gid,
context->ipc.perm_mode);
}
break; }
case AUDIT_MQ_OPEN:
audit_log_format(ab,
2011-07-26 09:26:10 +00:00
"oflag=0x%x mode=%#ho mq_flags=0x%lx mq_maxmsg=%ld "
"mq_msgsize=%ld mq_curmsgs=%ld",
context->mq_open.oflag, context->mq_open.mode,
context->mq_open.attr.mq_flags,
context->mq_open.attr.mq_maxmsg,
context->mq_open.attr.mq_msgsize,
context->mq_open.attr.mq_curmsgs);
break;
case AUDIT_MQ_SENDRECV:
audit_log_format(ab,
"mqdes=%d msg_len=%zd msg_prio=%u "
"abs_timeout_sec=%lld abs_timeout_nsec=%ld",
context->mq_sendrecv.mqdes,
context->mq_sendrecv.msg_len,
context->mq_sendrecv.msg_prio,
(long long) context->mq_sendrecv.abs_timeout.tv_sec,
context->mq_sendrecv.abs_timeout.tv_nsec);
break;
case AUDIT_MQ_NOTIFY:
audit_log_format(ab, "mqdes=%d sigev_signo=%d",
context->mq_notify.mqdes,
context->mq_notify.sigev_signo);
break;
case AUDIT_MQ_GETSETATTR: {
struct mq_attr *attr = &context->mq_getsetattr.mqstat;
audit_log_format(ab,
"mqdes=%d mq_flags=0x%lx mq_maxmsg=%ld mq_msgsize=%ld "
"mq_curmsgs=%ld ",
context->mq_getsetattr.mqdes,
attr->mq_flags, attr->mq_maxmsg,
attr->mq_msgsize, attr->mq_curmsgs);
break; }
case AUDIT_CAPSET:
audit_log_format(ab, "pid=%d", context->capset.pid);
audit_log_cap(ab, "cap_pi", &context->capset.cap.inheritable);
audit_log_cap(ab, "cap_pp", &context->capset.cap.permitted);
audit_log_cap(ab, "cap_pe", &context->capset.cap.effective);
audit: add ambient capabilities to CAPSET and BPRM_FCAPS records Capabilities were augmented to include ambient capabilities in v4.3 commit 58319057b784 ("capabilities: ambient capabilities"). Add ambient capabilities to the audit BPRM_FCAPS and CAPSET records. The record contains fields "old_pp", "old_pi", "old_pe", "new_pp", "new_pi", "new_pe" so in keeping with the previous record normalizations, change the "new_*" variants to simply drop the "new_" prefix. A sample of the replaced BPRM_FCAPS record: RAW: type=BPRM_FCAPS msg=audit(1491468034.252:237): fver=2 fp=0000000000200000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 old_pa=0000000000000000 pp=0000000000200000 pi=0000000000000000 pe=0000000000200000 pa=0000000000000000 INTERPRET: type=BPRM_FCAPS msg=audit(04/06/2017 04:40:34.252:237): fver=2 fp=sys_admin fi=none fe=chown old_pp=none old_pi=none old_pe=none old_pa=none pp=sys_admin pi=none pe=sys_admin pa=none A sample of the replaced CAPSET record: RAW: type=CAPSET msg=audit(1491469502.371:242): pid=833 cap_pi=0000003fffffffff cap_pp=0000003fffffffff cap_pe=0000003fffffffff cap_pa=0000000000000000 INTERPRET: type=CAPSET msg=audit(04/06/2017 05:05:02.371:242) : pid=833 cap_pi=chown,dac_override,dac_read_search,fowner,fsetid,kill, setgid,setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource,sys_time, sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pp=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pe=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pa=none See: https://github.com/linux-audit/audit-kernel/issues/40 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Serge Hallyn <serge@hallyn.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-04-07 14:17:27 +00:00
audit_log_cap(ab, "cap_pa", &context->capset.cap.ambient);
break;
case AUDIT_MMAP:
audit_log_format(ab, "fd=%d flags=0x%x", context->mmap.fd,
context->mmap.flags);
break;
audit: add OPENAT2 record to list "how" info Since the openat2(2) syscall uses a struct open_how pointer to communicate its parameters they are not usefully recorded by the audit SYSCALL record's four existing arguments. Add a new audit record type OPENAT2 that reports the parameters in its third argument, struct open_how with fields oflag, mode and resolve. The new record in the context of an event would look like: time->Wed Mar 17 16:28:53 2021 type=PROCTITLE msg=audit(1616012933.531:184): proctitle= 73797363616C6C735F66696C652F6F70656E617432002F746D702F61756469742D 7465737473756974652D737641440066696C652D6F70656E617432 type=PATH msg=audit(1616012933.531:184): item=1 name="file-openat2" inode=29 dev=00:1f mode=0100600 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmp_t:s0 nametype=CREATE cap_fp=0 cap_fi=0 cap_fe=0 cap_fver=0 cap_frootid=0 type=PATH msg=audit(1616012933.531:184): item=0 name="/root/rgb/git/audit-testsuite/tests" inode=25 dev=00:1f mode=040700 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmp_t:s0 nametype=PARENT cap_fp=0 cap_fi=0 cap_fe=0 cap_fver=0 cap_frootid=0 type=CWD msg=audit(1616012933.531:184): cwd="/root/rgb/git/audit-testsuite/tests" type=OPENAT2 msg=audit(1616012933.531:184): oflag=0100302 mode=0600 resolve=0xa type=SYSCALL msg=audit(1616012933.531:184): arch=c000003e syscall=437 success=yes exit=4 a0=3 a1=7ffe315f1c53 a2=7ffe315f1550 a3=18 items=2 ppid=528 pid=540 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=ttyS0 ses=1 comm="openat2" exe="/root/rgb/git/audit-testsuite/tests/syscalls_file/openat2" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key="testsuite-1616012933-bjAUcEPO" Link: https://lore.kernel.org/r/d23fbb89186754487850367224b060e26f9b7181.1621363275.git.rgb@redhat.com Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Christian Brauner <christian.brauner@ubuntu.com> [PM: tweak subject, wrap example, move AUDIT_OPENAT2 to 1337] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-05-19 20:00:22 +00:00
case AUDIT_OPENAT2:
audit_log_format(ab, "oflag=0%llo mode=0%llo resolve=0x%llx",
context->openat2.flags,
context->openat2.mode,
context->openat2.resolve);
break;
case AUDIT_EXECVE:
audit_log_execve_info(context, &ab);
break;
case AUDIT_KERN_MODULE:
audit_log_format(ab, "name=");
if (context->module.name) {
audit_log_untrustedstring(ab, context->module.name);
} else
audit_log_format(ab, "(null)");
break;
case AUDIT_TIME_ADJNTPVAL:
case AUDIT_TIME_INJOFFSET:
/* this call deviates from the rest, eating the buffer */
audit_log_time(context, &ab);
break;
}
audit_log_end(ab);
}
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
static inline int audit_proctitle_rtrim(char *proctitle, int len)
{
char *end = proctitle + len - 1;
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
while (end > proctitle && !isprint(*end))
end--;
/* catch the case where proctitle is only 1 non-print character */
len = end - proctitle + 1;
len -= isprint(proctitle[len-1]) == 0;
return len;
}
/*
* audit_log_name - produce AUDIT_PATH record from struct audit_names
* @context: audit_context for the task
* @n: audit_names structure with reportable details
* @path: optional path to report instead of audit_names->name
* @record_num: record number to report when handling a list of names
* @call_panic: optional pointer to int that will be updated if secid fails
*/
static void audit_log_name(struct audit_context *context, struct audit_names *n,
const struct path *path, int record_num, int *call_panic)
{
struct audit_buffer *ab;
ab = audit_log_start(context, GFP_KERNEL, AUDIT_PATH);
if (!ab)
return;
audit_log_format(ab, "item=%d", record_num);
if (path)
audit_log_d_path(ab, " name=", path);
else if (n->name) {
switch (n->name_len) {
case AUDIT_NAME_FULL:
/* log the full path */
audit_log_format(ab, " name=");
audit_log_untrustedstring(ab, n->name->name);
break;
case 0:
/* name was specified as a relative path and the
* directory component is the cwd
*/
audit: trigger accompanying records when no rules present When there are no audit rules registered, mandatory records (config, etc.) are missing their accompanying records (syscall, proctitle, etc.). This is due to audit context dummy set on syscall entry based on absence of rules that signals that no other records are to be printed. Clear the dummy bit if any record is generated, open coding this in audit_log_start(). The proctitle context and dummy checks are pointless since the proctitle record will not be printed if no syscall records are printed. The fds array is reset to -1 after the first syscall to indicate it isn't valid any more, but was never set to -1 when the context was allocated to indicate it wasn't yet valid. Check ctx->pwd in audit_log_name(). The audit_inode* functions can be called without going through getname_flags() or getname_kernel() that sets audit_names and cwd, so set the cwd in audit_alloc_name() if it has not already been done so due to audit_names being valid and purge all other audit_getcwd() calls. Revert the LSM dump_common_audit_data() LSM_AUDIT_DATA_* cases from the ghak96 patch since they are no longer necessary due to cwd coverage in audit_alloc_name(). Thanks to bauen1 <j2468h@googlemail.com> for reporting LSM situations in which context->cwd is not valid, inadvertantly fixed by the ghak96 patch. Please see upstream github issue https://github.com/linux-audit/audit-kernel/issues/120 This is also related to upstream github issue https://github.com/linux-audit/audit-kernel/issues/96 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-09-22 12:44:50 +00:00
if (context->pwd.dentry && context->pwd.mnt)
audit_log_d_path(ab, " name=", &context->pwd);
else
audit_log_format(ab, " name=(null)");
break;
default:
/* log the name's directory component */
audit_log_format(ab, " name=");
audit_log_n_untrustedstring(ab, n->name->name,
n->name_len);
}
} else
audit_log_format(ab, " name=(null)");
if (n->ino != AUDIT_INO_UNSET)
audit_log_format(ab, " inode=%lu dev=%02x:%02x mode=%#ho ouid=%u ogid=%u rdev=%02x:%02x",
n->ino,
MAJOR(n->dev),
MINOR(n->dev),
n->mode,
from_kuid(&init_user_ns, n->uid),
from_kgid(&init_user_ns, n->gid),
MAJOR(n->rdev),
MINOR(n->rdev));
if (n->osid != 0) {
char *ctx = NULL;
u32 len;
if (security_secid_to_secctx(
n->osid, &ctx, &len)) {
audit_log_format(ab, " osid=%u", n->osid);
if (call_panic)
*call_panic = 2;
} else {
audit_log_format(ab, " obj=%s", ctx);
security_release_secctx(ctx, len);
}
}
/* log the audit_names record type */
switch (n->type) {
case AUDIT_TYPE_NORMAL:
audit_log_format(ab, " nametype=NORMAL");
break;
case AUDIT_TYPE_PARENT:
audit_log_format(ab, " nametype=PARENT");
break;
case AUDIT_TYPE_CHILD_DELETE:
audit_log_format(ab, " nametype=DELETE");
break;
case AUDIT_TYPE_CHILD_CREATE:
audit_log_format(ab, " nametype=CREATE");
break;
default:
audit_log_format(ab, " nametype=UNKNOWN");
break;
}
audit_log_fcaps(ab, n);
audit_log_end(ab);
}
static void audit_log_proctitle(void)
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
{
int res;
char *buf;
char *msg = "(null)";
int len = strlen(msg);
struct audit_context *context = audit_context();
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
struct audit_buffer *ab;
ab = audit_log_start(context, GFP_KERNEL, AUDIT_PROCTITLE);
if (!ab)
return; /* audit_panic or being filtered */
audit_log_format(ab, "proctitle=");
/* Not cached */
if (!context->proctitle.value) {
buf = kmalloc(MAX_PROCTITLE_AUDIT_LEN, GFP_KERNEL);
if (!buf)
goto out;
/* Historically called this from procfs naming */
res = get_cmdline(current, buf, MAX_PROCTITLE_AUDIT_LEN);
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
if (res == 0) {
kfree(buf);
goto out;
}
res = audit_proctitle_rtrim(buf, res);
if (res == 0) {
kfree(buf);
goto out;
}
context->proctitle.value = buf;
context->proctitle.len = res;
}
msg = context->proctitle.value;
len = context->proctitle.len;
out:
audit_log_n_untrustedstring(ab, msg, len);
audit_log_end(ab);
}
/**
* audit_log_uring - generate a AUDIT_URINGOP record
* @ctx: the audit context
*/
static void audit_log_uring(struct audit_context *ctx)
{
struct audit_buffer *ab;
const struct cred *cred;
ab = audit_log_start(ctx, GFP_ATOMIC, AUDIT_URINGOP);
if (!ab)
return;
cred = current_cred();
audit_log_format(ab, "uring_op=%d", ctx->uring_op);
if (ctx->return_valid != AUDITSC_INVALID)
audit_log_format(ab, " success=%s exit=%ld",
(ctx->return_valid == AUDITSC_SUCCESS ?
"yes" : "no"),
ctx->return_code);
audit_log_format(ab,
" items=%d"
" ppid=%d pid=%d uid=%u gid=%u euid=%u suid=%u"
" fsuid=%u egid=%u sgid=%u fsgid=%u",
ctx->name_count,
task_ppid_nr(current), task_tgid_nr(current),
from_kuid(&init_user_ns, cred->uid),
from_kgid(&init_user_ns, cred->gid),
from_kuid(&init_user_ns, cred->euid),
from_kuid(&init_user_ns, cred->suid),
from_kuid(&init_user_ns, cred->fsuid),
from_kgid(&init_user_ns, cred->egid),
from_kgid(&init_user_ns, cred->sgid),
from_kgid(&init_user_ns, cred->fsgid));
audit_log_task_context(ab);
audit_log_key(ab, ctx->filterkey);
audit_log_end(ab);
}
static void audit_log_exit(void)
{
int i, call_panic = 0;
struct audit_context *context = audit_context();
struct audit_buffer *ab;
struct audit_aux_data *aux;
struct audit_names *n;
context->personality = current->personality;
switch (context->context) {
case AUDIT_CTX_SYSCALL:
ab = audit_log_start(context, GFP_KERNEL, AUDIT_SYSCALL);
if (!ab)
return;
audit_log_format(ab, "arch=%x syscall=%d",
context->arch, context->major);
if (context->personality != PER_LINUX)
audit_log_format(ab, " per=%lx", context->personality);
if (context->return_valid != AUDITSC_INVALID)
audit_log_format(ab, " success=%s exit=%ld",
(context->return_valid == AUDITSC_SUCCESS ?
"yes" : "no"),
context->return_code);
audit_log_format(ab,
" a0=%lx a1=%lx a2=%lx a3=%lx items=%d",
context->argv[0],
context->argv[1],
context->argv[2],
context->argv[3],
context->name_count);
audit_log_task_info(ab);
audit_log_key(ab, context->filterkey);
audit_log_end(ab);
break;
case AUDIT_CTX_URING:
audit_log_uring(context);
break;
default:
BUG();
break;
}
for (aux = context->aux; aux; aux = aux->next) {
ab = audit_log_start(context, GFP_KERNEL, aux->type);
if (!ab)
continue; /* audit_panic has been called */
switch (aux->type) {
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
case AUDIT_BPRM_FCAPS: {
struct audit_aux_data_bprm_fcaps *axs = (void *)aux;
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
audit_log_format(ab, "fver=%x", axs->fcap_ver);
audit_log_cap(ab, "fp", &axs->fcap.permitted);
audit_log_cap(ab, "fi", &axs->fcap.inheritable);
audit_log_format(ab, " fe=%d", axs->fcap.fE);
audit_log_cap(ab, "old_pp", &axs->old_pcap.permitted);
audit_log_cap(ab, "old_pi", &axs->old_pcap.inheritable);
audit_log_cap(ab, "old_pe", &axs->old_pcap.effective);
audit: add ambient capabilities to CAPSET and BPRM_FCAPS records Capabilities were augmented to include ambient capabilities in v4.3 commit 58319057b784 ("capabilities: ambient capabilities"). Add ambient capabilities to the audit BPRM_FCAPS and CAPSET records. The record contains fields "old_pp", "old_pi", "old_pe", "new_pp", "new_pi", "new_pe" so in keeping with the previous record normalizations, change the "new_*" variants to simply drop the "new_" prefix. A sample of the replaced BPRM_FCAPS record: RAW: type=BPRM_FCAPS msg=audit(1491468034.252:237): fver=2 fp=0000000000200000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 old_pa=0000000000000000 pp=0000000000200000 pi=0000000000000000 pe=0000000000200000 pa=0000000000000000 INTERPRET: type=BPRM_FCAPS msg=audit(04/06/2017 04:40:34.252:237): fver=2 fp=sys_admin fi=none fe=chown old_pp=none old_pi=none old_pe=none old_pa=none pp=sys_admin pi=none pe=sys_admin pa=none A sample of the replaced CAPSET record: RAW: type=CAPSET msg=audit(1491469502.371:242): pid=833 cap_pi=0000003fffffffff cap_pp=0000003fffffffff cap_pe=0000003fffffffff cap_pa=0000000000000000 INTERPRET: type=CAPSET msg=audit(04/06/2017 05:05:02.371:242) : pid=833 cap_pi=chown,dac_override,dac_read_search,fowner,fsetid,kill, setgid,setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource,sys_time, sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pp=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pe=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pa=none See: https://github.com/linux-audit/audit-kernel/issues/40 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Serge Hallyn <serge@hallyn.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-04-07 14:17:27 +00:00
audit_log_cap(ab, "old_pa", &axs->old_pcap.ambient);
audit_log_cap(ab, "pp", &axs->new_pcap.permitted);
audit_log_cap(ab, "pi", &axs->new_pcap.inheritable);
audit_log_cap(ab, "pe", &axs->new_pcap.effective);
audit_log_cap(ab, "pa", &axs->new_pcap.ambient);
audit_log_format(ab, " frootid=%d",
from_kuid(&init_user_ns,
axs->fcap.rootid));
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
break; }
}
audit_log_end(ab);
}
if (context->type)
show_special(context, &call_panic);
if (context->fds[0] >= 0) {
ab = audit_log_start(context, GFP_KERNEL, AUDIT_FD_PAIR);
if (ab) {
audit_log_format(ab, "fd0=%d fd1=%d",
context->fds[0], context->fds[1]);
audit_log_end(ab);
}
}
if (context->sockaddr_len) {
ab = audit_log_start(context, GFP_KERNEL, AUDIT_SOCKADDR);
if (ab) {
audit_log_format(ab, "saddr=");
audit_log_n_hex(ab, (void *)context->sockaddr,
context->sockaddr_len);
audit_log_end(ab);
}
}
for (aux = context->aux_pids; aux; aux = aux->next) {
struct audit_aux_data_pids *axs = (void *)aux;
for (i = 0; i < axs->pid_count; i++)
if (audit_log_pid_context(context, axs->target_pid[i],
axs->target_auid[i],
axs->target_uid[i],
axs->target_sessionid[i],
axs->target_sid[i],
axs->target_comm[i]))
call_panic = 1;
}
if (context->target_pid &&
audit_log_pid_context(context, context->target_pid,
context->target_auid, context->target_uid,
context->target_sessionid,
context->target_sid, context->target_comm))
call_panic = 1;
if (context->pwd.dentry && context->pwd.mnt) {
ab = audit_log_start(context, GFP_KERNEL, AUDIT_CWD);
if (ab) {
audit_log_d_path(ab, "cwd=", &context->pwd);
audit_log_end(ab);
}
}
i = 0;
audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reported-by: Jiri Jaburek <jjaburek@redhat.com> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-08 22:59:36 +00:00
list_for_each_entry(n, &context->names_list, list) {
if (n->hidden)
continue;
audit_log_name(context, n, NULL, i++, &call_panic);
audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reported-by: Jiri Jaburek <jjaburek@redhat.com> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-08 22:59:36 +00:00
}
if (context->context == AUDIT_CTX_SYSCALL)
audit_log_proctitle();
audit: Audit proc/<pid>/cmdline aka proctitle During an audit event, cache and print the value of the process's proctitle value (proc/<pid>/cmdline). This is useful in situations where processes are started via fork'd virtual machines where the comm field is incorrect. Often times, setting the comm field still is insufficient as the comm width is not very wide and most virtual machine "package names" do not fit. Also, during execution, many threads have their comm field set as well. By tying it back to the global cmdline value for the process, audit records will be more complete in systems with these properties. An example of where this is useful and applicable is in the realm of Android. With Android, their is no fork/exec for VM instances. The bare, preloaded Dalvik VM listens for a fork and specialize request. When this request comes in, the VM forks, and the loads the specific application (specializing). This was done to take advantage of COW and to not require a load of basic packages by the VM on very app spawn. When this spawn occurs, the package name is set via setproctitle() and shows up in procfs. Many of these package names are longer then 16 bytes, the historical width of task->comm. Having the cmdline in the audit records will couple the application back to the record directly. Also, on my Debian development box, some audit records were more useful then what was printed under comm. The cached proctitle is tied to the life-cycle of the audit_context structure and is built on demand. Proctitle is controllable by userspace, and thus should not be trusted. It is meant as an aid to assist in debugging. The proctitle event is emitted during syscall audits, and can be filtered with auditctl. Example: type=AVC msg=audit(1391217013.924:386): avc: denied { getattr } for pid=1971 comm="mkdir" name="/" dev="selinuxfs" ino=1 scontext=system_u:system_r:consolekit_t:s0-s0:c0.c255 tcontext=system_u:object_r:security_t:s0 tclass=filesystem type=SYSCALL msg=audit(1391217013.924:386): arch=c000003e syscall=137 success=yes exit=0 a0=7f019dfc8bd7 a1=7fffa6aed2c0 a2=fffffffffff4bd25 a3=7fffa6aed050 items=0 ppid=1967 pid=1971 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="mkdir" exe="/bin/mkdir" subj=system_u:system_r:consolekit_t:s0-s0:c0.c255 key=(null) type=UNKNOWN[1327] msg=audit(1391217013.924:386): proctitle=6D6B646972002D70002F7661722F72756E2F636F6E736F6C65 Acked-by: Steve Grubb <sgrubb@redhat.com> (wrt record formating) Signed-off-by: William Roberts <wroberts@tresys.com> Signed-off-by: Eric Paris <eparis@redhat.com>
2014-02-11 18:12:01 +00:00
/* Send end of event record to help user space know we are finished */
ab = audit_log_start(context, GFP_KERNEL, AUDIT_EOE);
if (ab)
audit_log_end(ab);
if (call_panic)
audit_panic("error in audit_log_exit()");
}
/**
* __audit_free - free a per-task audit context
* @tsk: task whose audit context block to free
*
* Called from copy_process, do_exit, and the io_uring code
*/
void __audit_free(struct task_struct *tsk)
{
struct audit_context *context = tsk->audit_context;
if (!context)
return;
/* this may generate CONFIG_CHANGE records */
if (!list_empty(&context->killed_trees))
audit_kill_trees(context);
/* We are called either by do_exit() or the fork() error handling code;
* in the former case tsk == current and in the latter tsk is a
* random task_struct that doesn't have any meaningful data we
* need to log via audit_log_exit().
*/
if (tsk == current && !context->dummy) {
context->return_valid = AUDITSC_INVALID;
context->return_code = 0;
if (context->context == AUDIT_CTX_SYSCALL) {
audit_filter_syscall(tsk, context);
audit_filter_inodes(tsk, context);
if (context->current_state == AUDIT_STATE_RECORD)
audit_log_exit();
} else if (context->context == AUDIT_CTX_URING) {
/* TODO: verify this case is real and valid */
audit_filter_uring(tsk, context);
audit_filter_inodes(tsk, context);
if (context->current_state == AUDIT_STATE_RECORD)
audit_log_uring(context);
}
}
audit_set_context(tsk, NULL);
audit_free_context(context);
}
/**
* audit_return_fixup - fixup the return codes in the audit_context
* @ctx: the audit_context
* @success: true/false value to indicate if the operation succeeded or not
* @code: operation return code
*
* We need to fixup the return code in the audit logs if the actual return
* codes are later going to be fixed by the arch specific signal handlers.
*/
static void audit_return_fixup(struct audit_context *ctx,
int success, long code)
{
/*
* This is actually a test for:
* (rc == ERESTARTSYS ) || (rc == ERESTARTNOINTR) ||
* (rc == ERESTARTNOHAND) || (rc == ERESTART_RESTARTBLOCK)
*
* but is faster than a bunch of ||
*/
if (unlikely(code <= -ERESTARTSYS) &&
(code >= -ERESTART_RESTARTBLOCK) &&
(code != -ENOIOCTLCMD))
ctx->return_code = -EINTR;
else
ctx->return_code = code;
ctx->return_valid = (success ? AUDITSC_SUCCESS : AUDITSC_FAILURE);
}
/**
* __audit_uring_entry - prepare the kernel task's audit context for io_uring
* @op: the io_uring opcode
*
* This is similar to audit_syscall_entry() but is intended for use by io_uring
* operations. This function should only ever be called from
* audit_uring_entry() as we rely on the audit context checking present in that
* function.
*/
void __audit_uring_entry(u8 op)
{
struct audit_context *ctx = audit_context();
if (ctx->state == AUDIT_STATE_DISABLED)
return;
/*
* NOTE: It's possible that we can be called from the process' context
* before it returns to userspace, and before audit_syscall_exit()
* is called. In this case there is not much to do, just record
* the io_uring details and return.
*/
ctx->uring_op = op;
if (ctx->context == AUDIT_CTX_SYSCALL)
return;
ctx->dummy = !audit_n_rules;
if (!ctx->dummy && ctx->state == AUDIT_STATE_BUILD)
ctx->prio = 0;
ctx->context = AUDIT_CTX_URING;
ctx->current_state = ctx->state;
ktime_get_coarse_real_ts64(&ctx->ctime);
}
/**
* __audit_uring_exit - wrap up the kernel task's audit context after io_uring
* @success: true/false value to indicate if the operation succeeded or not
* @code: operation return code
*
* This is similar to audit_syscall_exit() but is intended for use by io_uring
* operations. This function should only ever be called from
* audit_uring_exit() as we rely on the audit context checking present in that
* function.
*/
void __audit_uring_exit(int success, long code)
{
struct audit_context *ctx = audit_context();
if (ctx->dummy) {
if (ctx->context != AUDIT_CTX_URING)
return;
goto out;
}
audit_return_fixup(ctx, success, code);
if (ctx->context == AUDIT_CTX_SYSCALL) {
/*
* NOTE: See the note in __audit_uring_entry() about the case
* where we may be called from process context before we
* return to userspace via audit_syscall_exit(). In this
* case we simply emit a URINGOP record and bail, the
* normal syscall exit handling will take care of
* everything else.
* It is also worth mentioning that when we are called,
* the current process creds may differ from the creds
* used during the normal syscall processing; keep that
* in mind if/when we move the record generation code.
*/
/*
* We need to filter on the syscall info here to decide if we
* should emit a URINGOP record. I know it seems odd but this
* solves the problem where users have a filter to block *all*
* syscall records in the "exit" filter; we want to preserve
* the behavior here.
*/
audit_filter_syscall(current, ctx);
if (ctx->current_state != AUDIT_STATE_RECORD)
audit_filter_uring(current, ctx);
audit_filter_inodes(current, ctx);
if (ctx->current_state != AUDIT_STATE_RECORD)
return;
audit_log_uring(ctx);
return;
}
/* this may generate CONFIG_CHANGE records */
if (!list_empty(&ctx->killed_trees))
audit_kill_trees(ctx);
/* run through both filters to ensure we set the filterkey properly */
audit_filter_uring(current, ctx);
audit_filter_inodes(current, ctx);
if (ctx->current_state != AUDIT_STATE_RECORD)
goto out;
audit_log_exit();
out:
audit_reset_context(ctx);
}
/**
* __audit_syscall_entry - fill in an audit record at syscall entry
* @major: major syscall type (function)
* @a1: additional syscall register 1
* @a2: additional syscall register 2
* @a3: additional syscall register 3
* @a4: additional syscall register 4
*
* Fill in audit context at syscall entry. This only happens if the
* audit context was created when the task was created and the state or
* filters demand the audit context be built. If the state from the
* per-task filter or from the per-syscall filter is AUDIT_STATE_RECORD,
* then the record will be written at syscall exit time (otherwise, it
* will only be written if another part of the kernel requests that it
* be written).
*/
void __audit_syscall_entry(int major, unsigned long a1, unsigned long a2,
unsigned long a3, unsigned long a4)
{
struct audit_context *context = audit_context();
enum audit_state state;
if (!audit_enabled || !context)
return;
WARN_ON(context->context != AUDIT_CTX_UNUSED);
WARN_ON(context->name_count);
if (context->context != AUDIT_CTX_UNUSED || context->name_count) {
audit_panic("unrecoverable error in audit_syscall_entry()");
return;
}
state = context->state;
if (state == AUDIT_STATE_DISABLED)
return;
context->dummy = !audit_n_rules;
if (!context->dummy && state == AUDIT_STATE_BUILD) {
context->prio = 0;
if (auditd_test_task(current))
return;
}
syscall_get_arch: add "struct task_struct *" argument This argument is required to extend the generic ptrace API with PTRACE_GET_SYSCALL_INFO request: syscall_get_arch() is going to be called from ptrace_request() along with syscall_get_nr(), syscall_get_arguments(), syscall_get_error(), and syscall_get_return_value() functions with a tracee as their argument. The primary intent is that the triple (audit_arch, syscall_nr, arg1..arg6) should describe what system call is being called and what its arguments are. Reverts: 5e937a9ae913 ("syscall_get_arch: remove useless function arguments") Reverts: 1002d94d3076 ("syscall.h: fix doc text for syscall_get_arch()") Reviewed-by: Andy Lutomirski <luto@kernel.org> # for x86 Reviewed-by: Palmer Dabbelt <palmer@sifive.com> Acked-by: Paul Moore <paul@paul-moore.com> Acked-by: Paul Burton <paul.burton@mips.com> # MIPS parts Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc) Acked-by: Kees Cook <keescook@chromium.org> # seccomp parts Acked-by: Mark Salter <msalter@redhat.com> # for the c6x bit Cc: Elvira Khabirova <lineprinter@altlinux.org> Cc: Eugene Syromyatnikov <esyr@redhat.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: x86@kernel.org Cc: linux-alpha@vger.kernel.org Cc: linux-snps-arc@lists.infradead.org Cc: linux-arm-kernel@lists.infradead.org Cc: linux-c6x-dev@linux-c6x.org Cc: uclinux-h8-devel@lists.sourceforge.jp Cc: linux-hexagon@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-m68k@lists.linux-m68k.org Cc: linux-mips@vger.kernel.org Cc: nios2-dev@lists.rocketboards.org Cc: openrisc@lists.librecores.org Cc: linux-parisc@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Cc: linux-riscv@lists.infradead.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: sparclinux@vger.kernel.org Cc: linux-um@lists.infradead.org Cc: linux-xtensa@linux-xtensa.org Cc: linux-arch@vger.kernel.org Cc: linux-audit@redhat.com Signed-off-by: Dmitry V. Levin <ldv@altlinux.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-03-17 23:30:18 +00:00
context->arch = syscall_get_arch(current);
context->major = major;
context->argv[0] = a1;
context->argv[1] = a2;
context->argv[2] = a3;
context->argv[3] = a4;
context->context = AUDIT_CTX_SYSCALL;
context->current_state = state;
ktime_get_coarse_real_ts64(&context->ctime);
}
/**
* __audit_syscall_exit - deallocate audit context after a system call
* @success: success value of the syscall
* @return_code: return value of the syscall
*
* Tear down after system call. If the audit context has been marked as
* auditable (either because of the AUDIT_STATE_RECORD state from
* filtering, or because some other part of the kernel wrote an audit
* message), then write out the syscall information. In call cases,
* free the names stored from getname().
*/
Audit: push audit success and retcode into arch ptrace.h The audit system previously expected arches calling to audit_syscall_exit to supply as arguments if the syscall was a success and what the return code was. Audit also provides a helper AUDITSC_RESULT which was supposed to simplify things by converting from negative retcodes to an audit internal magic value stating success or failure. This helper was wrong and could indicate that a valid pointer returned to userspace was a failed syscall. The fix is to fix the layering foolishness. We now pass audit_syscall_exit a struct pt_reg and it in turns calls back into arch code to collect the return value and to determine if the syscall was a success or failure. We also define a generic is_syscall_success() macro which determines success/failure based on if the value is < -MAX_ERRNO. This works for arches like x86 which do not use a separate mechanism to indicate syscall failure. We make both the is_syscall_success() and regs_return_value() static inlines instead of macros. The reason is because the audit function must take a void* for the regs. (uml calls theirs struct uml_pt_regs instead of just struct pt_regs so audit_syscall_exit can't take a struct pt_regs). Since the audit function takes a void* we need to use static inlines to cast it back to the arch correct structure to dereference it. The other major change is that on some arches, like ia64, MIPS and ppc, we change regs_return_value() to give us the negative value on syscall failure. THE only other user of this macro, kretprobe_example.c, won't notice and it makes the value signed consistently for the audit functions across all archs. In arch/sh/kernel/ptrace_64.c I see that we were using regs[9] in the old audit code as the return value. But the ptrace_64.h code defined the macro regs_return_value() as regs[3]. I have no idea which one is correct, but this patch now uses the regs_return_value() function, so it now uses regs[3]. For powerpc we previously used regs->result but now use the regs_return_value() function which uses regs->gprs[3]. regs->gprs[3] is always positive so the regs_return_value(), much like ia64 makes it negative before calling the audit code when appropriate. Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: H. Peter Anvin <hpa@zytor.com> [for x86 portion] Acked-by: Tony Luck <tony.luck@intel.com> [for ia64] Acked-by: Richard Weinberger <richard@nod.at> [for uml] Acked-by: David S. Miller <davem@davemloft.net> [for sparc] Acked-by: Ralf Baechle <ralf@linux-mips.org> [for mips] Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> [for ppc]
2012-01-03 19:23:06 +00:00
void __audit_syscall_exit(int success, long return_code)
{
struct audit_context *context = audit_context();
if (!context || context->dummy ||
context->context != AUDIT_CTX_SYSCALL)
goto out;
/* this may generate CONFIG_CHANGE records */
if (!list_empty(&context->killed_trees))
audit_kill_trees(context);
audit_return_fixup(context, success, return_code);
/* run through both filters to ensure we set the filterkey properly */
audit_filter_syscall(current, context);
audit_filter_inodes(current, context);
if (context->current_state != AUDIT_STATE_RECORD)
goto out;
audit_log_exit();
out:
audit_reset_context(context);
}
static inline void handle_one(const struct inode *inode)
{
struct audit_context *context;
struct audit_tree_refs *p;
struct audit_chunk *chunk;
int count;
if (likely(!inode->i_fsnotify_marks))
return;
context = audit_context();
p = context->trees;
count = context->tree_count;
rcu_read_lock();
chunk = audit_tree_lookup(inode);
rcu_read_unlock();
if (!chunk)
return;
if (likely(put_tree_ref(context, chunk)))
return;
if (unlikely(!grow_tree_refs(context))) {
pr_warn("out of memory, audit has lost a tree reference\n");
audit_set_auditable(context);
audit_put_chunk(chunk);
unroll_tree_refs(context, p, count);
return;
}
put_tree_ref(context, chunk);
}
static void handle_path(const struct dentry *dentry)
{
struct audit_context *context;
struct audit_tree_refs *p;
const struct dentry *d, *parent;
struct audit_chunk *drop;
unsigned long seq;
int count;
context = audit_context();
p = context->trees;
count = context->tree_count;
retry:
drop = NULL;
d = dentry;
rcu_read_lock();
seq = read_seqbegin(&rename_lock);
for (;;) {
struct inode *inode = d_backing_inode(d);
if (inode && unlikely(inode->i_fsnotify_marks)) {
struct audit_chunk *chunk;
chunk = audit_tree_lookup(inode);
if (chunk) {
if (unlikely(!put_tree_ref(context, chunk))) {
drop = chunk;
break;
}
}
}
parent = d->d_parent;
if (parent == d)
break;
d = parent;
}
if (unlikely(read_seqretry(&rename_lock, seq) || drop)) { /* in this order */
rcu_read_unlock();
if (!drop) {
/* just a race with rename */
unroll_tree_refs(context, p, count);
goto retry;
}
audit_put_chunk(drop);
if (grow_tree_refs(context)) {
/* OK, got more space */
unroll_tree_refs(context, p, count);
goto retry;
}
/* too bad */
pr_warn("out of memory, audit has lost a tree reference\n");
unroll_tree_refs(context, p, count);
audit_set_auditable(context);
return;
}
rcu_read_unlock();
}
static struct audit_names *audit_alloc_name(struct audit_context *context,
unsigned char type)
{
struct audit_names *aname;
if (context->name_count < AUDIT_NAMES) {
aname = &context->preallocated_names[context->name_count];
memset(aname, 0, sizeof(*aname));
} else {
aname = kzalloc(sizeof(*aname), GFP_NOFS);
if (!aname)
return NULL;
aname->should_free = true;
}
aname->ino = AUDIT_INO_UNSET;
aname->type = type;
list_add_tail(&aname->list, &context->names_list);
context->name_count++;
audit: trigger accompanying records when no rules present When there are no audit rules registered, mandatory records (config, etc.) are missing their accompanying records (syscall, proctitle, etc.). This is due to audit context dummy set on syscall entry based on absence of rules that signals that no other records are to be printed. Clear the dummy bit if any record is generated, open coding this in audit_log_start(). The proctitle context and dummy checks are pointless since the proctitle record will not be printed if no syscall records are printed. The fds array is reset to -1 after the first syscall to indicate it isn't valid any more, but was never set to -1 when the context was allocated to indicate it wasn't yet valid. Check ctx->pwd in audit_log_name(). The audit_inode* functions can be called without going through getname_flags() or getname_kernel() that sets audit_names and cwd, so set the cwd in audit_alloc_name() if it has not already been done so due to audit_names being valid and purge all other audit_getcwd() calls. Revert the LSM dump_common_audit_data() LSM_AUDIT_DATA_* cases from the ghak96 patch since they are no longer necessary due to cwd coverage in audit_alloc_name(). Thanks to bauen1 <j2468h@googlemail.com> for reporting LSM situations in which context->cwd is not valid, inadvertantly fixed by the ghak96 patch. Please see upstream github issue https://github.com/linux-audit/audit-kernel/issues/120 This is also related to upstream github issue https://github.com/linux-audit/audit-kernel/issues/96 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2020-09-22 12:44:50 +00:00
if (!context->pwd.dentry)
get_fs_pwd(current->fs, &context->pwd);
return aname;
}
/**
* __audit_reusename - fill out filename with info from existing entry
* @uptr: userland ptr to pathname
*
* Search the audit_names list for the current audit context. If there is an
* existing entry with a matching "uptr" then return the filename
* associated with that audit_name. If not, return NULL.
*/
struct filename *
__audit_reusename(const __user char *uptr)
{
struct audit_context *context = audit_context();
struct audit_names *n;
list_for_each_entry(n, &context->names_list, list) {
if (!n->name)
continue;
if (n->name->uptr == uptr) {
audit,io_uring: io_uring openat triggers audit reference count underflow An io_uring openat operation can update an audit reference count from multiple threads resulting in the call trace below. A call to io_uring_submit() with a single openat op with a flag of IOSQE_ASYNC results in the following reference count updates. These first part of the system call performs two increments that do not race. do_syscall_64() __do_sys_io_uring_enter() io_submit_sqes() io_openat_prep() __io_openat_prep() getname() getname_flags() /* update 1 (increment) */ __audit_getname() /* update 2 (increment) */ The openat op is queued to an io_uring worker thread which starts the opportunity for a race. The system call exit performs one decrement. do_syscall_64() syscall_exit_to_user_mode() syscall_exit_to_user_mode_prepare() __audit_syscall_exit() audit_reset_context() putname() /* update 3 (decrement) */ The io_uring worker thread performs one increment and two decrements. These updates can race with the system call decrement. io_wqe_worker() io_worker_handle_work() io_wq_submit_work() io_issue_sqe() io_openat() io_openat2() do_filp_open() path_openat() __audit_inode() /* update 4 (increment) */ putname() /* update 5 (decrement) */ __audit_uring_exit() audit_reset_context() putname() /* update 6 (decrement) */ The fix is to change the refcnt member of struct audit_names from int to atomic_t. kernel BUG at fs/namei.c:262! Call Trace: ... ? putname+0x68/0x70 audit_reset_context.part.0.constprop.0+0xe1/0x300 __audit_uring_exit+0xda/0x1c0 io_issue_sqe+0x1f3/0x450 ? lock_timer_base+0x3b/0xd0 io_wq_submit_work+0x8d/0x2b0 ? __try_to_del_timer_sync+0x67/0xa0 io_worker_handle_work+0x17c/0x2b0 io_wqe_worker+0x10a/0x350 Cc: stable@vger.kernel.org Link: https://lore.kernel.org/lkml/MW2PR2101MB1033FFF044A258F84AEAA584F1C9A@MW2PR2101MB1033.namprd21.prod.outlook.com/ Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Signed-off-by: Dan Clash <daclash@linux.microsoft.com> Link: https://lore.kernel.org/r/20231012215518.GA4048@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Reviewed-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-10-12 21:55:18 +00:00
atomic_inc(&n->name->refcnt);
return n->name;
}
}
return NULL;
}
/**
* __audit_getname - add a name to the list
* @name: name to add
*
* Add a name to the list of audit names for this context.
* Called from fs/namei.c:getname().
*/
void __audit_getname(struct filename *name)
{
struct audit_context *context = audit_context();
struct audit_names *n;
if (context->context == AUDIT_CTX_UNUSED)
return;
n = audit_alloc_name(context, AUDIT_TYPE_UNKNOWN);
if (!n)
return;
n->name = name;
n->name_len = AUDIT_NAME_FULL;
name->aname = n;
audit,io_uring: io_uring openat triggers audit reference count underflow An io_uring openat operation can update an audit reference count from multiple threads resulting in the call trace below. A call to io_uring_submit() with a single openat op with a flag of IOSQE_ASYNC results in the following reference count updates. These first part of the system call performs two increments that do not race. do_syscall_64() __do_sys_io_uring_enter() io_submit_sqes() io_openat_prep() __io_openat_prep() getname() getname_flags() /* update 1 (increment) */ __audit_getname() /* update 2 (increment) */ The openat op is queued to an io_uring worker thread which starts the opportunity for a race. The system call exit performs one decrement. do_syscall_64() syscall_exit_to_user_mode() syscall_exit_to_user_mode_prepare() __audit_syscall_exit() audit_reset_context() putname() /* update 3 (decrement) */ The io_uring worker thread performs one increment and two decrements. These updates can race with the system call decrement. io_wqe_worker() io_worker_handle_work() io_wq_submit_work() io_issue_sqe() io_openat() io_openat2() do_filp_open() path_openat() __audit_inode() /* update 4 (increment) */ putname() /* update 5 (decrement) */ __audit_uring_exit() audit_reset_context() putname() /* update 6 (decrement) */ The fix is to change the refcnt member of struct audit_names from int to atomic_t. kernel BUG at fs/namei.c:262! Call Trace: ... ? putname+0x68/0x70 audit_reset_context.part.0.constprop.0+0xe1/0x300 __audit_uring_exit+0xda/0x1c0 io_issue_sqe+0x1f3/0x450 ? lock_timer_base+0x3b/0xd0 io_wq_submit_work+0x8d/0x2b0 ? __try_to_del_timer_sync+0x67/0xa0 io_worker_handle_work+0x17c/0x2b0 io_wqe_worker+0x10a/0x350 Cc: stable@vger.kernel.org Link: https://lore.kernel.org/lkml/MW2PR2101MB1033FFF044A258F84AEAA584F1C9A@MW2PR2101MB1033.namprd21.prod.outlook.com/ Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Signed-off-by: Dan Clash <daclash@linux.microsoft.com> Link: https://lore.kernel.org/r/20231012215518.GA4048@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Reviewed-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-10-12 21:55:18 +00:00
atomic_inc(&name->refcnt);
}
static inline int audit_copy_fcaps(struct audit_names *name,
const struct dentry *dentry)
{
struct cpu_vfs_cap_data caps;
int rc;
if (!dentry)
return 0;
rc = get_vfs_caps_from_disk(&nop_mnt_idmap, dentry, &caps);
if (rc)
return rc;
name->fcap.permitted = caps.permitted;
name->fcap.inheritable = caps.inheritable;
name->fcap.fE = !!(caps.magic_etc & VFS_CAP_FLAGS_EFFECTIVE);
name->fcap.rootid = caps.rootid;
name->fcap_ver = (caps.magic_etc & VFS_CAP_REVISION_MASK) >>
VFS_CAP_REVISION_SHIFT;
return 0;
}
/* Copy inode data into an audit_names. */
static void audit_copy_inode(struct audit_names *name,
const struct dentry *dentry,
struct inode *inode, unsigned int flags)
{
name->ino = inode->i_ino;
name->dev = inode->i_sb->s_dev;
name->mode = inode->i_mode;
name->uid = inode->i_uid;
name->gid = inode->i_gid;
name->rdev = inode->i_rdev;
security_inode_getsecid(inode, &name->osid);
if (flags & AUDIT_INODE_NOEVAL) {
name->fcap_ver = -1;
return;
}
audit_copy_fcaps(name, dentry);
}
/**
* __audit_inode - store the inode and device from a lookup
* @name: name being audited
* @dentry: dentry being audited
audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reported-by: Jiri Jaburek <jjaburek@redhat.com> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-08 22:59:36 +00:00
* @flags: attributes for this particular entry
*/
void __audit_inode(struct filename *name, const struct dentry *dentry,
audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reported-by: Jiri Jaburek <jjaburek@redhat.com> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-08 22:59:36 +00:00
unsigned int flags)
{
struct audit_context *context = audit_context();
struct inode *inode = d_backing_inode(dentry);
struct audit_names *n;
audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reported-by: Jiri Jaburek <jjaburek@redhat.com> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-08 22:59:36 +00:00
bool parent = flags & AUDIT_INODE_PARENT;
struct audit_entry *e;
struct list_head *list = &audit_filter_list[AUDIT_FILTER_FS];
int i;
if (context->context == AUDIT_CTX_UNUSED)
return;
rcu_read_lock();
list_for_each_entry_rcu(e, list, list) {
for (i = 0; i < e->rule.field_count; i++) {
struct audit_field *f = &e->rule.fields[i];
if (f->type == AUDIT_FSTYPE
&& audit_comparator(inode->i_sb->s_magic,
f->op, f->val)
&& e->rule.action == AUDIT_NEVER) {
rcu_read_unlock();
return;
}
}
}
rcu_read_unlock();
if (!name)
goto out_alloc;
/*
* If we have a pointer to an audit_names entry already, then we can
* just use it directly if the type is correct.
*/
n = name->aname;
if (n) {
if (parent) {
if (n->type == AUDIT_TYPE_PARENT ||
n->type == AUDIT_TYPE_UNKNOWN)
goto out;
} else {
if (n->type != AUDIT_TYPE_PARENT)
goto out;
}
}
list_for_each_entry_reverse(n, &context->names_list, list) {
if (n->ino) {
/* valid inode number, use that for the comparison */
if (n->ino != inode->i_ino ||
n->dev != inode->i_sb->s_dev)
continue;
} else if (n->name) {
/* inode number has not been set, check the name */
if (strcmp(n->name->name, name->name))
continue;
} else
/* no inode and no name (?!) ... this is odd ... */
continue;
/* match the correct record type */
if (parent) {
if (n->type == AUDIT_TYPE_PARENT ||
n->type == AUDIT_TYPE_UNKNOWN)
goto out;
} else {
if (n->type != AUDIT_TYPE_PARENT)
goto out;
}
}
out_alloc:
audit: correctly record file names with different path name types There is a problem with the audit system when multiple audit records are created for the same path, each with a different path name type. The root cause of the problem is in __audit_inode() when an exact match (both the path name and path name type) is not found for a path name record; the existing code creates a new path name record, but it never sets the path name in this record, leaving it NULL. This patch corrects this problem by assigning the path name to these newly created records. There are many ways to reproduce this problem, but one of the easiest is the following (assuming auditd is running): # mkdir /root/tmp/test # touch /root/tmp/test/567 # auditctl -a always,exit -F dir=/root/tmp/test # touch /root/tmp/test/567 Afterwards, or while the commands above are running, check the audit log and pay special attention to the PATH records. A faulty kernel will display something like the following for the file creation: type=SYSCALL msg=audit(1416957442.025:93): arch=c000003e syscall=2 success=yes exit=3 ... comm="touch" exe="/usr/bin/touch" type=CWD msg=audit(1416957442.025:93): cwd="/root/tmp" type=PATH msg=audit(1416957442.025:93): item=0 name="test/" inode=401409 ... nametype=PARENT type=PATH msg=audit(1416957442.025:93): item=1 name=(null) inode=393804 ... nametype=NORMAL type=PATH msg=audit(1416957442.025:93): item=2 name=(null) inode=393804 ... nametype=NORMAL While a patched kernel will show the following: type=SYSCALL msg=audit(1416955786.566:89): arch=c000003e syscall=2 success=yes exit=3 ... comm="touch" exe="/usr/bin/touch" type=CWD msg=audit(1416955786.566:89): cwd="/root/tmp" type=PATH msg=audit(1416955786.566:89): item=0 name="test/" inode=401409 ... nametype=PARENT type=PATH msg=audit(1416955786.566:89): item=1 name="test/567" inode=393804 ... nametype=NORMAL This issue was brought up by a number of people, but special credit should go to hujianyang@huawei.com for reporting the problem along with an explanation of the problem and a patch. While the original patch did have some problems (see the archive link below), it did demonstrate the problem and helped kickstart the fix presented here. * https://lkml.org/lkml/2014/9/5/66 Reported-by: hujianyang <hujianyang@huawei.com> Signed-off-by: Paul Moore <pmoore@redhat.com> Acked-by: Richard Guy Briggs <rgb@redhat.com>
2014-12-22 17:27:39 +00:00
/* unable to find an entry with both a matching name and type */
n = audit_alloc_name(context, AUDIT_TYPE_UNKNOWN);
if (!n)
return;
if (name) {
n->name = name;
audit,io_uring: io_uring openat triggers audit reference count underflow An io_uring openat operation can update an audit reference count from multiple threads resulting in the call trace below. A call to io_uring_submit() with a single openat op with a flag of IOSQE_ASYNC results in the following reference count updates. These first part of the system call performs two increments that do not race. do_syscall_64() __do_sys_io_uring_enter() io_submit_sqes() io_openat_prep() __io_openat_prep() getname() getname_flags() /* update 1 (increment) */ __audit_getname() /* update 2 (increment) */ The openat op is queued to an io_uring worker thread which starts the opportunity for a race. The system call exit performs one decrement. do_syscall_64() syscall_exit_to_user_mode() syscall_exit_to_user_mode_prepare() __audit_syscall_exit() audit_reset_context() putname() /* update 3 (decrement) */ The io_uring worker thread performs one increment and two decrements. These updates can race with the system call decrement. io_wqe_worker() io_worker_handle_work() io_wq_submit_work() io_issue_sqe() io_openat() io_openat2() do_filp_open() path_openat() __audit_inode() /* update 4 (increment) */ putname() /* update 5 (decrement) */ __audit_uring_exit() audit_reset_context() putname() /* update 6 (decrement) */ The fix is to change the refcnt member of struct audit_names from int to atomic_t. kernel BUG at fs/namei.c:262! Call Trace: ... ? putname+0x68/0x70 audit_reset_context.part.0.constprop.0+0xe1/0x300 __audit_uring_exit+0xda/0x1c0 io_issue_sqe+0x1f3/0x450 ? lock_timer_base+0x3b/0xd0 io_wq_submit_work+0x8d/0x2b0 ? __try_to_del_timer_sync+0x67/0xa0 io_worker_handle_work+0x17c/0x2b0 io_wqe_worker+0x10a/0x350 Cc: stable@vger.kernel.org Link: https://lore.kernel.org/lkml/MW2PR2101MB1033FFF044A258F84AEAA584F1C9A@MW2PR2101MB1033.namprd21.prod.outlook.com/ Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Signed-off-by: Dan Clash <daclash@linux.microsoft.com> Link: https://lore.kernel.org/r/20231012215518.GA4048@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Reviewed-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-10-12 21:55:18 +00:00
atomic_inc(&name->refcnt);
}
audit: correctly record file names with different path name types There is a problem with the audit system when multiple audit records are created for the same path, each with a different path name type. The root cause of the problem is in __audit_inode() when an exact match (both the path name and path name type) is not found for a path name record; the existing code creates a new path name record, but it never sets the path name in this record, leaving it NULL. This patch corrects this problem by assigning the path name to these newly created records. There are many ways to reproduce this problem, but one of the easiest is the following (assuming auditd is running): # mkdir /root/tmp/test # touch /root/tmp/test/567 # auditctl -a always,exit -F dir=/root/tmp/test # touch /root/tmp/test/567 Afterwards, or while the commands above are running, check the audit log and pay special attention to the PATH records. A faulty kernel will display something like the following for the file creation: type=SYSCALL msg=audit(1416957442.025:93): arch=c000003e syscall=2 success=yes exit=3 ... comm="touch" exe="/usr/bin/touch" type=CWD msg=audit(1416957442.025:93): cwd="/root/tmp" type=PATH msg=audit(1416957442.025:93): item=0 name="test/" inode=401409 ... nametype=PARENT type=PATH msg=audit(1416957442.025:93): item=1 name=(null) inode=393804 ... nametype=NORMAL type=PATH msg=audit(1416957442.025:93): item=2 name=(null) inode=393804 ... nametype=NORMAL While a patched kernel will show the following: type=SYSCALL msg=audit(1416955786.566:89): arch=c000003e syscall=2 success=yes exit=3 ... comm="touch" exe="/usr/bin/touch" type=CWD msg=audit(1416955786.566:89): cwd="/root/tmp" type=PATH msg=audit(1416955786.566:89): item=0 name="test/" inode=401409 ... nametype=PARENT type=PATH msg=audit(1416955786.566:89): item=1 name="test/567" inode=393804 ... nametype=NORMAL This issue was brought up by a number of people, but special credit should go to hujianyang@huawei.com for reporting the problem along with an explanation of the problem and a patch. While the original patch did have some problems (see the archive link below), it did demonstrate the problem and helped kickstart the fix presented here. * https://lkml.org/lkml/2014/9/5/66 Reported-by: hujianyang <hujianyang@huawei.com> Signed-off-by: Paul Moore <pmoore@redhat.com> Acked-by: Richard Guy Briggs <rgb@redhat.com>
2014-12-22 17:27:39 +00:00
out:
if (parent) {
n->name_len = n->name ? parent_len(n->name->name) : AUDIT_NAME_FULL;
n->type = AUDIT_TYPE_PARENT;
audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record The old audit PATH records for mq_open looked like this: type=PATH msg=audit(1366282323.982:869): item=1 name=(null) inode=6777 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282323.982:869): item=0 name="test_mq" inode=26732 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 ...with the audit related changes that went into 3.7, they now look like this: type=PATH msg=audit(1366282236.776:3606): item=2 name=(null) inode=66655 dev=00:0c mode=0100700 ouid=0 ogid=0 rdev=00:00 obj=staff_u:object_r:user_tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=1 name=(null) inode=6926 dev=00:0c mode=041777 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:tmpfs_t:s15:c0.c1023 type=PATH msg=audit(1366282236.776:3606): item=0 name="test_mq" Both of these look wrong to me. As Steve Grubb pointed out: "What we need is 1 PATH record that identifies the MQ. The other PATH records probably should not be there." Fix it to record the mq root as a parent, and flag it such that it should be hidden from view when the names are logged, since the root of the mq filesystem isn't terribly interesting. With this change, we get a single PATH record that looks more like this: type=PATH msg=audit(1368021604.836:484): item=0 name="test_mq" inode=16914 dev=00:0c mode=0100644 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmpfs_t:s0 In order to do this, a new audit_inode_parent_hidden() function is added. If we do it this way, then we avoid having the existing callers of audit_inode needing to do any sort of flag conversion if auditing is inactive. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reported-by: Jiri Jaburek <jjaburek@redhat.com> Cc: Steve Grubb <sgrubb@redhat.com> Cc: Eric Paris <eparis@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-07-08 22:59:36 +00:00
if (flags & AUDIT_INODE_HIDDEN)
n->hidden = true;
} else {
n->name_len = AUDIT_NAME_FULL;
n->type = AUDIT_TYPE_NORMAL;
}
handle_path(dentry);
audit_copy_inode(n, dentry, inode, flags & AUDIT_INODE_NOEVAL);
}
void __audit_file(const struct file *file)
{
__audit_inode(NULL, file->f_path.dentry, 0);
}
/**
* __audit_inode_child - collect inode info for created/removed objects
* @parent: inode of dentry parent
* @dentry: dentry being audited
* @type: AUDIT_TYPE_* value that we're looking for
*
* For syscalls that create or remove filesystem objects, audit_inode
* can only collect information for the filesystem object's parent.
* This call updates the audit context with the child's information.
* Syscalls that create a new filesystem object must be hooked after
* the object is created. Syscalls that remove a filesystem object
* must be hooked prior, in order to capture the target inode during
* unsuccessful attempts.
*/
void __audit_inode_child(struct inode *parent,
const struct dentry *dentry,
const unsigned char type)
{
struct audit_context *context = audit_context();
struct inode *inode = d_backing_inode(dentry);
const struct qstr *dname = &dentry->d_name;
struct audit_names *n, *found_parent = NULL, *found_child = NULL;
struct audit_entry *e;
struct list_head *list = &audit_filter_list[AUDIT_FILTER_FS];
int i;
if (context->context == AUDIT_CTX_UNUSED)
return;
rcu_read_lock();
list_for_each_entry_rcu(e, list, list) {
for (i = 0; i < e->rule.field_count; i++) {
struct audit_field *f = &e->rule.fields[i];
if (f->type == AUDIT_FSTYPE
&& audit_comparator(parent->i_sb->s_magic,
f->op, f->val)
&& e->rule.action == AUDIT_NEVER) {
rcu_read_unlock();
return;
}
}
}
rcu_read_unlock();
if (inode)
handle_one(inode);
/* look for a parent entry first */
list_for_each_entry(n, &context->names_list, list) {
if (!n->name ||
(n->type != AUDIT_TYPE_PARENT &&
n->type != AUDIT_TYPE_UNKNOWN))
continue;
if (n->ino == parent->i_ino && n->dev == parent->i_sb->s_dev &&
!audit_compare_dname_path(dname,
n->name->name, n->name_len)) {
if (n->type == AUDIT_TYPE_UNKNOWN)
n->type = AUDIT_TYPE_PARENT;
found_parent = n;
break;
}
}
audit: fix possible soft lockup in __audit_inode_child() Tracefs or debugfs maybe cause hundreds to thousands of PATH records, too many PATH records maybe cause soft lockup. For example: 1. CONFIG_KASAN=y && CONFIG_PREEMPTION=n 2. auditctl -a exit,always -S open -k key 3. sysctl -w kernel.watchdog_thresh=5 4. mkdir /sys/kernel/debug/tracing/instances/test There may be a soft lockup as follows: watchdog: BUG: soft lockup - CPU#45 stuck for 7s! [mkdir:15498] Kernel panic - not syncing: softlockup: hung tasks Call trace: dump_backtrace+0x0/0x30c show_stack+0x20/0x30 dump_stack+0x11c/0x174 panic+0x27c/0x494 watchdog_timer_fn+0x2bc/0x390 __run_hrtimer+0x148/0x4fc __hrtimer_run_queues+0x154/0x210 hrtimer_interrupt+0x2c4/0x760 arch_timer_handler_phys+0x48/0x60 handle_percpu_devid_irq+0xe0/0x340 __handle_domain_irq+0xbc/0x130 gic_handle_irq+0x78/0x460 el1_irq+0xb8/0x140 __audit_inode_child+0x240/0x7bc tracefs_create_file+0x1b8/0x2a0 trace_create_file+0x18/0x50 event_create_dir+0x204/0x30c __trace_add_new_event+0xac/0x100 event_trace_add_tracer+0xa0/0x130 trace_array_create_dir+0x60/0x140 trace_array_create+0x1e0/0x370 instance_mkdir+0x90/0xd0 tracefs_syscall_mkdir+0x68/0xa0 vfs_mkdir+0x21c/0x34c do_mkdirat+0x1b4/0x1d4 __arm64_sys_mkdirat+0x4c/0x60 el0_svc_common.constprop.0+0xa8/0x240 do_el0_svc+0x8c/0xc0 el0_svc+0x20/0x30 el0_sync_handler+0xb0/0xb4 el0_sync+0x160/0x180 Therefore, we add cond_resched() to __audit_inode_child() to fix it. Fixes: 5195d8e217a7 ("audit: dynamically allocate audit_names when not enough space is in the names array") Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2023-08-08 12:14:35 +00:00
cond_resched();
/* is there a matching child entry? */
list_for_each_entry(n, &context->names_list, list) {
/* can only match entries that have a name */
if (!n->name ||
(n->type != type && n->type != AUDIT_TYPE_UNKNOWN))
continue;
if (!strcmp(dname->name, n->name->name) ||
!audit_compare_dname_path(dname, n->name->name,
found_parent ?
found_parent->name_len :
AUDIT_NAME_FULL)) {
if (n->type == AUDIT_TYPE_UNKNOWN)
n->type = type;
found_child = n;
break;
}
}
if (!found_parent) {
/* create a new, "anonymous" parent record */
n = audit_alloc_name(context, AUDIT_TYPE_PARENT);
if (!n)
return;
audit_copy_inode(n, NULL, parent, 0);
}
if (!found_child) {
found_child = audit_alloc_name(context, type);
if (!found_child)
return;
/* Re-use the name belonging to the slot for a matching parent
* directory. All names for this context are relinquished in
* audit_free_names() */
if (found_parent) {
found_child->name = found_parent->name;
found_child->name_len = AUDIT_NAME_FULL;
audit,io_uring: io_uring openat triggers audit reference count underflow An io_uring openat operation can update an audit reference count from multiple threads resulting in the call trace below. A call to io_uring_submit() with a single openat op with a flag of IOSQE_ASYNC results in the following reference count updates. These first part of the system call performs two increments that do not race. do_syscall_64() __do_sys_io_uring_enter() io_submit_sqes() io_openat_prep() __io_openat_prep() getname() getname_flags() /* update 1 (increment) */ __audit_getname() /* update 2 (increment) */ The openat op is queued to an io_uring worker thread which starts the opportunity for a race. The system call exit performs one decrement. do_syscall_64() syscall_exit_to_user_mode() syscall_exit_to_user_mode_prepare() __audit_syscall_exit() audit_reset_context() putname() /* update 3 (decrement) */ The io_uring worker thread performs one increment and two decrements. These updates can race with the system call decrement. io_wqe_worker() io_worker_handle_work() io_wq_submit_work() io_issue_sqe() io_openat() io_openat2() do_filp_open() path_openat() __audit_inode() /* update 4 (increment) */ putname() /* update 5 (decrement) */ __audit_uring_exit() audit_reset_context() putname() /* update 6 (decrement) */ The fix is to change the refcnt member of struct audit_names from int to atomic_t. kernel BUG at fs/namei.c:262! Call Trace: ... ? putname+0x68/0x70 audit_reset_context.part.0.constprop.0+0xe1/0x300 __audit_uring_exit+0xda/0x1c0 io_issue_sqe+0x1f3/0x450 ? lock_timer_base+0x3b/0xd0 io_wq_submit_work+0x8d/0x2b0 ? __try_to_del_timer_sync+0x67/0xa0 io_worker_handle_work+0x17c/0x2b0 io_wqe_worker+0x10a/0x350 Cc: stable@vger.kernel.org Link: https://lore.kernel.org/lkml/MW2PR2101MB1033FFF044A258F84AEAA584F1C9A@MW2PR2101MB1033.namprd21.prod.outlook.com/ Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Signed-off-by: Dan Clash <daclash@linux.microsoft.com> Link: https://lore.kernel.org/r/20231012215518.GA4048@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Reviewed-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Christian Brauner <brauner@kernel.org>
2023-10-12 21:55:18 +00:00
atomic_inc(&found_child->name->refcnt);
}
}
if (inode)
audit_copy_inode(found_child, dentry, inode, 0);
else
found_child->ino = AUDIT_INO_UNSET;
}
EXPORT_SYMBOL_GPL(__audit_inode_child);
/**
* auditsc_get_stamp - get local copies of audit_context values
* @ctx: audit_context for the task
* @t: timespec64 to store time recorded in the audit_context
* @serial: serial value that is recorded in the audit_context
*
* Also sets the context as auditable.
*/
int auditsc_get_stamp(struct audit_context *ctx,
struct timespec64 *t, unsigned int *serial)
{
if (ctx->context == AUDIT_CTX_UNUSED)
return 0;
if (!ctx->serial)
ctx->serial = audit_serial();
t->tv_sec = ctx->ctime.tv_sec;
t->tv_nsec = ctx->ctime.tv_nsec;
*serial = ctx->serial;
if (!ctx->prio) {
ctx->prio = 1;
ctx->current_state = AUDIT_STATE_RECORD;
}
return 1;
}
/**
* __audit_mq_open - record audit data for a POSIX MQ open
* @oflag: open flag
* @mode: mode bits
* @attr: queue attributes
*
*/
2011-07-26 09:26:10 +00:00
void __audit_mq_open(int oflag, umode_t mode, struct mq_attr *attr)
{
struct audit_context *context = audit_context();
if (attr)
memcpy(&context->mq_open.attr, attr, sizeof(struct mq_attr));
else
memset(&context->mq_open.attr, 0, sizeof(struct mq_attr));
context->mq_open.oflag = oflag;
context->mq_open.mode = mode;
context->type = AUDIT_MQ_OPEN;
}
/**
* __audit_mq_sendrecv - record audit data for a POSIX MQ timed send/receive
* @mqdes: MQ descriptor
* @msg_len: Message length
* @msg_prio: Message priority
* @abs_timeout: Message timeout in absolute time
*
*/
void __audit_mq_sendrecv(mqd_t mqdes, size_t msg_len, unsigned int msg_prio,
const struct timespec64 *abs_timeout)
{
struct audit_context *context = audit_context();
struct timespec64 *p = &context->mq_sendrecv.abs_timeout;
if (abs_timeout)
memcpy(p, abs_timeout, sizeof(*p));
else
memset(p, 0, sizeof(*p));
context->mq_sendrecv.mqdes = mqdes;
context->mq_sendrecv.msg_len = msg_len;
context->mq_sendrecv.msg_prio = msg_prio;
context->type = AUDIT_MQ_SENDRECV;
}
/**
* __audit_mq_notify - record audit data for a POSIX MQ notify
* @mqdes: MQ descriptor
* @notification: Notification event
*
*/
void __audit_mq_notify(mqd_t mqdes, const struct sigevent *notification)
{
struct audit_context *context = audit_context();
if (notification)
context->mq_notify.sigev_signo = notification->sigev_signo;
else
context->mq_notify.sigev_signo = 0;
context->mq_notify.mqdes = mqdes;
context->type = AUDIT_MQ_NOTIFY;
}
/**
* __audit_mq_getsetattr - record audit data for a POSIX MQ get/set attribute
* @mqdes: MQ descriptor
* @mqstat: MQ flags
*
*/
void __audit_mq_getsetattr(mqd_t mqdes, struct mq_attr *mqstat)
{
struct audit_context *context = audit_context();
context->mq_getsetattr.mqdes = mqdes;
context->mq_getsetattr.mqstat = *mqstat;
context->type = AUDIT_MQ_GETSETATTR;
}
/**
* __audit_ipc_obj - record audit data for ipc object
[PATCH] Rework of IPC auditing 1) The audit_ipc_perms() function has been split into two different functions: - audit_ipc_obj() - audit_ipc_set_perm() There's a key shift here... The audit_ipc_obj() collects the uid, gid, mode, and SElinux context label of the current ipc object. This audit_ipc_obj() hook is now found in several places. Most notably, it is hooked in ipcperms(), which is called in various places around the ipc code permforming a MAC check. Additionally there are several places where *checkid() is used to validate that an operation is being performed on a valid object while not necessarily having a nearby ipcperms() call. In these locations, audit_ipc_obj() is called to ensure that the information is captured by the audit system. The audit_set_new_perm() function is called any time the permissions on the ipc object changes. In this case, the NEW permissions are recorded (and note that an audit_ipc_obj() call exists just a few lines before each instance). 2) Support for an AUDIT_IPC_SET_PERM audit message type. This allows for separate auxiliary audit records for normal operations on an IPC object and permissions changes. Note that the same struct audit_aux_data_ipcctl is used and populated, however there are separate audit_log_format statements based on the type of the message. Finally, the AUDIT_IPC block of code in audit_free_aux() was extended to handle aux messages of this new type. No more mem leaks I hope ;-) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2006-04-02 21:07:33 +00:00
* @ipcp: ipc permissions
*
*/
void __audit_ipc_obj(struct kern_ipc_perm *ipcp)
[PATCH] Rework of IPC auditing 1) The audit_ipc_perms() function has been split into two different functions: - audit_ipc_obj() - audit_ipc_set_perm() There's a key shift here... The audit_ipc_obj() collects the uid, gid, mode, and SElinux context label of the current ipc object. This audit_ipc_obj() hook is now found in several places. Most notably, it is hooked in ipcperms(), which is called in various places around the ipc code permforming a MAC check. Additionally there are several places where *checkid() is used to validate that an operation is being performed on a valid object while not necessarily having a nearby ipcperms() call. In these locations, audit_ipc_obj() is called to ensure that the information is captured by the audit system. The audit_set_new_perm() function is called any time the permissions on the ipc object changes. In this case, the NEW permissions are recorded (and note that an audit_ipc_obj() call exists just a few lines before each instance). 2) Support for an AUDIT_IPC_SET_PERM audit message type. This allows for separate auxiliary audit records for normal operations on an IPC object and permissions changes. Note that the same struct audit_aux_data_ipcctl is used and populated, however there are separate audit_log_format statements based on the type of the message. Finally, the AUDIT_IPC block of code in audit_free_aux() was extended to handle aux messages of this new type. No more mem leaks I hope ;-) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2006-04-02 21:07:33 +00:00
{
struct audit_context *context = audit_context();
context->ipc.uid = ipcp->uid;
context->ipc.gid = ipcp->gid;
context->ipc.mode = ipcp->mode;
context->ipc.has_perm = 0;
security_ipc_getsecid(ipcp, &context->ipc.osid);
context->type = AUDIT_IPC;
[PATCH] Rework of IPC auditing 1) The audit_ipc_perms() function has been split into two different functions: - audit_ipc_obj() - audit_ipc_set_perm() There's a key shift here... The audit_ipc_obj() collects the uid, gid, mode, and SElinux context label of the current ipc object. This audit_ipc_obj() hook is now found in several places. Most notably, it is hooked in ipcperms(), which is called in various places around the ipc code permforming a MAC check. Additionally there are several places where *checkid() is used to validate that an operation is being performed on a valid object while not necessarily having a nearby ipcperms() call. In these locations, audit_ipc_obj() is called to ensure that the information is captured by the audit system. The audit_set_new_perm() function is called any time the permissions on the ipc object changes. In this case, the NEW permissions are recorded (and note that an audit_ipc_obj() call exists just a few lines before each instance). 2) Support for an AUDIT_IPC_SET_PERM audit message type. This allows for separate auxiliary audit records for normal operations on an IPC object and permissions changes. Note that the same struct audit_aux_data_ipcctl is used and populated, however there are separate audit_log_format statements based on the type of the message. Finally, the AUDIT_IPC block of code in audit_free_aux() was extended to handle aux messages of this new type. No more mem leaks I hope ;-) Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2006-04-02 21:07:33 +00:00
}
/**
* __audit_ipc_set_perm - record audit data for new ipc permissions
* @qbytes: msgq bytes
* @uid: msgq user id
* @gid: msgq group id
* @mode: msgq mode (permissions)
*
* Called only after audit_ipc_obj().
*/
void __audit_ipc_set_perm(unsigned long qbytes, uid_t uid, gid_t gid, umode_t mode)
{
struct audit_context *context = audit_context();
context->ipc.qbytes = qbytes;
context->ipc.perm_uid = uid;
context->ipc.perm_gid = gid;
context->ipc.perm_mode = mode;
context->ipc.has_perm = 1;
}
void __audit_bprm(struct linux_binprm *bprm)
{
struct audit_context *context = audit_context();
context->type = AUDIT_EXECVE;
context->execve.argc = bprm->argc;
}
/**
* __audit_socketcall - record audit data for sys_socketcall
* @nargs: number of args, which should not be more than AUDITSC_ARGS.
* @args: args array
*
*/
int __audit_socketcall(int nargs, unsigned long *args)
{
struct audit_context *context = audit_context();
if (nargs <= 0 || nargs > AUDITSC_ARGS || !args)
return -EINVAL;
context->type = AUDIT_SOCKETCALL;
context->socketcall.nargs = nargs;
memcpy(context->socketcall.args, args, nargs * sizeof(unsigned long));
return 0;
}
/**
* __audit_fd_pair - record audit data for pipe and socketpair
* @fd1: the first file descriptor
* @fd2: the second file descriptor
*
*/
void __audit_fd_pair(int fd1, int fd2)
{
struct audit_context *context = audit_context();
context->fds[0] = fd1;
context->fds[1] = fd2;
}
/**
* __audit_sockaddr - record audit data for sys_bind, sys_connect, sys_sendto
* @len: data length in user space
* @a: data address in kernel space
*
* Returns 0 for success or NULL context or < 0 on error.
*/
int __audit_sockaddr(int len, void *a)
{
struct audit_context *context = audit_context();
if (!context->sockaddr) {
void *p = kmalloc(sizeof(struct sockaddr_storage), GFP_KERNEL);
if (!p)
return -ENOMEM;
context->sockaddr = p;
}
context->sockaddr_len = len;
memcpy(context->sockaddr, a, len);
return 0;
}
void __audit_ptrace(struct task_struct *t)
{
struct audit_context *context = audit_context();
context->target_pid = task_tgid_nr(t);
context->target_auid = audit_get_loginuid(t);
context->target_uid = task_uid(t);
context->target_sessionid = audit_get_sessionid(t);
security_task_getsecid_obj(t, &context->target_sid);
memcpy(context->target_comm, t->comm, TASK_COMM_LEN);
}
/**
* audit_signal_info_syscall - record signal info for syscalls
* @t: task being signaled
*
* If the audit subsystem is being terminated, record the task (pid)
* and uid that is doing that.
*/
int audit_signal_info_syscall(struct task_struct *t)
{
struct audit_aux_data_pids *axp;
struct audit_context *ctx = audit_context();
kuid_t t_uid = task_uid(t);
if (!audit_signals || audit_dummy_context())
return 0;
/* optimize the common case by putting first signal recipient directly
* in audit_context */
if (!ctx->target_pid) {
ctx->target_pid = task_tgid_nr(t);
ctx->target_auid = audit_get_loginuid(t);
ctx->target_uid = t_uid;
ctx->target_sessionid = audit_get_sessionid(t);
security_task_getsecid_obj(t, &ctx->target_sid);
memcpy(ctx->target_comm, t->comm, TASK_COMM_LEN);
return 0;
}
axp = (void *)ctx->aux_pids;
if (!axp || axp->pid_count == AUDIT_AUX_PIDS) {
axp = kzalloc(sizeof(*axp), GFP_ATOMIC);
if (!axp)
return -ENOMEM;
axp->d.type = AUDIT_OBJ_PID;
axp->d.next = ctx->aux_pids;
ctx->aux_pids = (void *)axp;
}
BUG_ON(axp->pid_count >= AUDIT_AUX_PIDS);
axp->target_pid[axp->pid_count] = task_tgid_nr(t);
axp->target_auid[axp->pid_count] = audit_get_loginuid(t);
axp->target_uid[axp->pid_count] = t_uid;
axp->target_sessionid[axp->pid_count] = audit_get_sessionid(t);
security_task_getsecid_obj(t, &axp->target_sid[axp->pid_count]);
memcpy(axp->target_comm[axp->pid_count], t->comm, TASK_COMM_LEN);
axp->pid_count++;
return 0;
}
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
/**
* __audit_log_bprm_fcaps - store information about a loading bprm and relevant fcaps
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
* @bprm: pointer to the bprm being processed
* @new: the proposed new credentials
* @old: the old credentials
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
*
* Simply check if the proc already has the caps given by the file and if not
* store the priv escalation info for later auditing at the end of the syscall
*
* -Eric
*/
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
int __audit_log_bprm_fcaps(struct linux_binprm *bprm,
const struct cred *new, const struct cred *old)
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
{
struct audit_aux_data_bprm_fcaps *ax;
struct audit_context *context = audit_context();
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
struct cpu_vfs_cap_data vcaps;
ax = kmalloc(sizeof(*ax), GFP_KERNEL);
if (!ax)
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
return -ENOMEM;
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
ax->d.type = AUDIT_BPRM_FCAPS;
ax->d.next = context->aux;
context->aux = (void *)ax;
get_vfs_caps_from_disk(&nop_mnt_idmap,
commoncap: handle idmapped mounts When interacting with user namespace and non-user namespace aware filesystem capabilities the vfs will perform various security checks to determine whether or not the filesystem capabilities can be used by the caller, whether they need to be removed and so on. The main infrastructure for this resides in the capability codepaths but they are called through the LSM security infrastructure even though they are not technically an LSM or optional. This extends the existing security hooks security_inode_removexattr(), security_inode_killpriv(), security_inode_getsecurity() to pass down the mount's user namespace and makes them aware of idmapped mounts. In order to actually get filesystem capabilities from disk the capability infrastructure exposes the get_vfs_caps_from_disk() helper. For user namespace aware filesystem capabilities a root uid is stored alongside the capabilities. In order to determine whether the caller can make use of the filesystem capability or whether it needs to be ignored it is translated according to the superblock's user namespace. If it can be translated to uid 0 according to that id mapping the caller can use the filesystem capabilities stored on disk. If we are accessing the inode that holds the filesystem capabilities through an idmapped mount we map the root uid according to the mount's user namespace. Afterwards the checks are identical to non-idmapped mounts: reading filesystem caps from disk enforces that the root uid associated with the filesystem capability must have a mapping in the superblock's user namespace and that the caller is either in the same user namespace or is a descendant of the superblock's user namespace. For filesystems that are mountable inside user namespace the caller can just mount the filesystem and won't usually need to idmap it. If they do want to idmap it they can create an idmapped mount and mark it with a user namespace they created and which is thus a descendant of s_user_ns. For filesystems that are not mountable inside user namespaces the descendant rule is trivially true because the s_user_ns will be the initial user namespace. If the initial user namespace is passed nothing changes so non-idmapped mounts will see identical behavior as before. Link: https://lore.kernel.org/r/20210121131959.646623-11-christian.brauner@ubuntu.com Cc: Christoph Hellwig <hch@lst.de> Cc: David Howells <dhowells@redhat.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: linux-fsdevel@vger.kernel.org Reviewed-by: Christoph Hellwig <hch@lst.de> Acked-by: James Morris <jamorris@linux.microsoft.com> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
2021-01-21 13:19:29 +00:00
bprm->file->f_path.dentry, &vcaps);
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
ax->fcap.permitted = vcaps.permitted;
ax->fcap.inheritable = vcaps.inheritable;
ax->fcap.fE = !!(vcaps.magic_etc & VFS_CAP_FLAGS_EFFECTIVE);
ax->fcap.rootid = vcaps.rootid;
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
ax->fcap_ver = (vcaps.magic_etc & VFS_CAP_REVISION_MASK) >> VFS_CAP_REVISION_SHIFT;
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
ax->old_pcap.permitted = old->cap_permitted;
ax->old_pcap.inheritable = old->cap_inheritable;
ax->old_pcap.effective = old->cap_effective;
audit: add ambient capabilities to CAPSET and BPRM_FCAPS records Capabilities were augmented to include ambient capabilities in v4.3 commit 58319057b784 ("capabilities: ambient capabilities"). Add ambient capabilities to the audit BPRM_FCAPS and CAPSET records. The record contains fields "old_pp", "old_pi", "old_pe", "new_pp", "new_pi", "new_pe" so in keeping with the previous record normalizations, change the "new_*" variants to simply drop the "new_" prefix. A sample of the replaced BPRM_FCAPS record: RAW: type=BPRM_FCAPS msg=audit(1491468034.252:237): fver=2 fp=0000000000200000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 old_pa=0000000000000000 pp=0000000000200000 pi=0000000000000000 pe=0000000000200000 pa=0000000000000000 INTERPRET: type=BPRM_FCAPS msg=audit(04/06/2017 04:40:34.252:237): fver=2 fp=sys_admin fi=none fe=chown old_pp=none old_pi=none old_pe=none old_pa=none pp=sys_admin pi=none pe=sys_admin pa=none A sample of the replaced CAPSET record: RAW: type=CAPSET msg=audit(1491469502.371:242): pid=833 cap_pi=0000003fffffffff cap_pp=0000003fffffffff cap_pe=0000003fffffffff cap_pa=0000000000000000 INTERPRET: type=CAPSET msg=audit(04/06/2017 05:05:02.371:242) : pid=833 cap_pi=chown,dac_override,dac_read_search,fowner,fsetid,kill, setgid,setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource,sys_time, sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pp=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pe=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pa=none See: https://github.com/linux-audit/audit-kernel/issues/40 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Serge Hallyn <serge@hallyn.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-04-07 14:17:27 +00:00
ax->old_pcap.ambient = old->cap_ambient;
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
ax->new_pcap.permitted = new->cap_permitted;
ax->new_pcap.inheritable = new->cap_inheritable;
ax->new_pcap.effective = new->cap_effective;
audit: add ambient capabilities to CAPSET and BPRM_FCAPS records Capabilities were augmented to include ambient capabilities in v4.3 commit 58319057b784 ("capabilities: ambient capabilities"). Add ambient capabilities to the audit BPRM_FCAPS and CAPSET records. The record contains fields "old_pp", "old_pi", "old_pe", "new_pp", "new_pi", "new_pe" so in keeping with the previous record normalizations, change the "new_*" variants to simply drop the "new_" prefix. A sample of the replaced BPRM_FCAPS record: RAW: type=BPRM_FCAPS msg=audit(1491468034.252:237): fver=2 fp=0000000000200000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 old_pa=0000000000000000 pp=0000000000200000 pi=0000000000000000 pe=0000000000200000 pa=0000000000000000 INTERPRET: type=BPRM_FCAPS msg=audit(04/06/2017 04:40:34.252:237): fver=2 fp=sys_admin fi=none fe=chown old_pp=none old_pi=none old_pe=none old_pa=none pp=sys_admin pi=none pe=sys_admin pa=none A sample of the replaced CAPSET record: RAW: type=CAPSET msg=audit(1491469502.371:242): pid=833 cap_pi=0000003fffffffff cap_pp=0000003fffffffff cap_pe=0000003fffffffff cap_pa=0000000000000000 INTERPRET: type=CAPSET msg=audit(04/06/2017 05:05:02.371:242) : pid=833 cap_pi=chown,dac_override,dac_read_search,fowner,fsetid,kill, setgid,setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource,sys_time, sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pp=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pe=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pa=none See: https://github.com/linux-audit/audit-kernel/issues/40 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Serge Hallyn <serge@hallyn.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-04-07 14:17:27 +00:00
ax->new_pcap.ambient = new->cap_ambient;
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
return 0;
Any time fcaps or a setuid app under SECURE_NOROOT is used to result in a non-zero pE we will crate a new audit record which contains the entire set of known information about the executable in question, fP, fI, fE, fversion and includes the process's pE, pI, pP. Before and after the bprm capability are applied. This record type will only be emitted from execve syscalls. an example of making ping use fcaps instead of setuid: setcap "cat_net_raw+pe" /bin/ping type=SYSCALL msg=audit(1225742021.015:236): arch=c000003e syscall=59 success=yes exit=0 a0=1457f30 a1=14606b0 a2=1463940 a3=321b770a70 items=2 ppid=2929 pid=2963 auid=0 uid=500 gid=500 euid=500 suid=500 fsuid=500 egid=500 sgid=500 fsgid=500 tty=pts0 ses=3 comm="ping" exe="/bin/ping" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key=(null) type=UNKNOWN[1321] msg=audit(1225742021.015:236): fver=2 fp=0000000000002000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 new_pp=0000000000002000 new_pi=0000000000000000 new_pe=0000000000002000 type=EXECVE msg=audit(1225742021.015:236): argc=2 a0="ping" a1="127.0.0.1" type=CWD msg=audit(1225742021.015:236): cwd="/home/test" type=PATH msg=audit(1225742021.015:236): item=0 name="/bin/ping" inode=49256 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ping_exec_t:s0 cap_fp=0000000000002000 cap_fe=1 cap_fver=2 type=PATH msg=audit(1225742021.015:236): item=1 name=(null) inode=507915 dev=fd:00 mode=0100755 ouid=0 ogid=0 rdev=00:00 obj=system_u:object_r:ld_so_t:s0 Signed-off-by: Eric Paris <eparis@redhat.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-11 10:48:18 +00:00
}
/**
* __audit_log_capset - store information about the arguments to the capset syscall
CRED: Inaugurate COW credentials Inaugurate copy-on-write credentials management. This uses RCU to manage the credentials pointer in the task_struct with respect to accesses by other tasks. A process may only modify its own credentials, and so does not need locking to access or modify its own credentials. A mutex (cred_replace_mutex) is added to the task_struct to control the effect of PTRACE_ATTACHED on credential calculations, particularly with respect to execve(). With this patch, the contents of an active credentials struct may not be changed directly; rather a new set of credentials must be prepared, modified and committed using something like the following sequence of events: struct cred *new = prepare_creds(); int ret = blah(new); if (ret < 0) { abort_creds(new); return ret; } return commit_creds(new); There are some exceptions to this rule: the keyrings pointed to by the active credentials may be instantiated - keyrings violate the COW rule as managing COW keyrings is tricky, given that it is possible for a task to directly alter the keys in a keyring in use by another task. To help enforce this, various pointers to sets of credentials, such as those in the task_struct, are declared const. The purpose of this is compile-time discouragement of altering credentials through those pointers. Once a set of credentials has been made public through one of these pointers, it may not be modified, except under special circumstances: (1) Its reference count may incremented and decremented. (2) The keyrings to which it points may be modified, but not replaced. The only safe way to modify anything else is to create a replacement and commit using the functions described in Documentation/credentials.txt (which will be added by a later patch). This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). This now prepares and commits credentials in various places in the security code rather than altering the current creds directly. (2) Temporary credential overrides. do_coredump() and sys_faccessat() now prepare their own credentials and temporarily override the ones currently on the acting thread, whilst preventing interference from other threads by holding cred_replace_mutex on the thread being dumped. This will be replaced in a future patch by something that hands down the credentials directly to the functions being called, rather than altering the task's objective credentials. (3) LSM interface. A number of functions have been changed, added or removed: (*) security_capset_check(), ->capset_check() (*) security_capset_set(), ->capset_set() Removed in favour of security_capset(). (*) security_capset(), ->capset() New. This is passed a pointer to the new creds, a pointer to the old creds and the proposed capability sets. It should fill in the new creds or return an error. All pointers, barring the pointer to the new creds, are now const. (*) security_bprm_apply_creds(), ->bprm_apply_creds() Changed; now returns a value, which will cause the process to be killed if it's an error. (*) security_task_alloc(), ->task_alloc_security() Removed in favour of security_prepare_creds(). (*) security_cred_free(), ->cred_free() New. Free security data attached to cred->security. (*) security_prepare_creds(), ->cred_prepare() New. Duplicate any security data attached to cred->security. (*) security_commit_creds(), ->cred_commit() New. Apply any security effects for the upcoming installation of new security by commit_creds(). (*) security_task_post_setuid(), ->task_post_setuid() Removed in favour of security_task_fix_setuid(). (*) security_task_fix_setuid(), ->task_fix_setuid() Fix up the proposed new credentials for setuid(). This is used by cap_set_fix_setuid() to implicitly adjust capabilities in line with setuid() changes. Changes are made to the new credentials, rather than the task itself as in security_task_post_setuid(). (*) security_task_reparent_to_init(), ->task_reparent_to_init() Removed. Instead the task being reparented to init is referred directly to init's credentials. NOTE! This results in the loss of some state: SELinux's osid no longer records the sid of the thread that forked it. (*) security_key_alloc(), ->key_alloc() (*) security_key_permission(), ->key_permission() Changed. These now take cred pointers rather than task pointers to refer to the security context. (4) sys_capset(). This has been simplified and uses less locking. The LSM functions it calls have been merged. (5) reparent_to_kthreadd(). This gives the current thread the same credentials as init by simply using commit_thread() to point that way. (6) __sigqueue_alloc() and switch_uid() __sigqueue_alloc() can't stop the target task from changing its creds beneath it, so this function gets a reference to the currently applicable user_struct which it then passes into the sigqueue struct it returns if successful. switch_uid() is now called from commit_creds(), and possibly should be folded into that. commit_creds() should take care of protecting __sigqueue_alloc(). (7) [sg]et[ug]id() and co and [sg]et_current_groups. The set functions now all use prepare_creds(), commit_creds() and abort_creds() to build and check a new set of credentials before applying it. security_task_set[ug]id() is called inside the prepared section. This guarantees that nothing else will affect the creds until we've finished. The calling of set_dumpable() has been moved into commit_creds(). Much of the functionality of set_user() has been moved into commit_creds(). The get functions all simply access the data directly. (8) security_task_prctl() and cap_task_prctl(). security_task_prctl() has been modified to return -ENOSYS if it doesn't want to handle a function, or otherwise return the return value directly rather than through an argument. Additionally, cap_task_prctl() now prepares a new set of credentials, even if it doesn't end up using it. (9) Keyrings. A number of changes have been made to the keyrings code: (a) switch_uid_keyring(), copy_keys(), exit_keys() and suid_keys() have all been dropped and built in to the credentials functions directly. They may want separating out again later. (b) key_alloc() and search_process_keyrings() now take a cred pointer rather than a task pointer to specify the security context. (c) copy_creds() gives a new thread within the same thread group a new thread keyring if its parent had one, otherwise it discards the thread keyring. (d) The authorisation key now points directly to the credentials to extend the search into rather pointing to the task that carries them. (e) Installing thread, process or session keyrings causes a new set of credentials to be created, even though it's not strictly necessary for process or session keyrings (they're shared). (10) Usermode helper. The usermode helper code now carries a cred struct pointer in its subprocess_info struct instead of a new session keyring pointer. This set of credentials is derived from init_cred and installed on the new process after it has been cloned. call_usermodehelper_setup() allocates the new credentials and call_usermodehelper_freeinfo() discards them if they haven't been used. A special cred function (prepare_usermodeinfo_creds()) is provided specifically for call_usermodehelper_setup() to call. call_usermodehelper_setkeys() adjusts the credentials to sport the supplied keyring as the new session keyring. (11) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) selinux_setprocattr() no longer does its check for whether the current ptracer can access processes with the new SID inside the lock that covers getting the ptracer's SID. Whilst this lock ensures that the check is done with the ptracer pinned, the result is only valid until the lock is released, so there's no point doing it inside the lock. (12) is_single_threaded(). This function has been extracted from selinux_setprocattr() and put into a file of its own in the lib/ directory as join_session_keyring() now wants to use it too. The code in SELinux just checked to see whether a task shared mm_structs with other tasks (CLONE_VM), but that isn't good enough. We really want to know if they're part of the same thread group (CLONE_THREAD). (13) nfsd. The NFS server daemon now has to use the COW credentials to set the credentials it is going to use. It really needs to pass the credentials down to the functions it calls, but it can't do that until other patches in this series have been applied. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: James Morris <jmorris@namei.org> Signed-off-by: James Morris <jmorris@namei.org>
2008-11-13 23:39:23 +00:00
* @new: the new credentials
* @old: the old (current) credentials
*
* Record the arguments userspace sent to sys_capset for later printing by the
* audit system if applicable
*/
void __audit_log_capset(const struct cred *new, const struct cred *old)
{
struct audit_context *context = audit_context();
context->capset.pid = task_tgid_nr(current);
context->capset.cap.effective = new->cap_effective;
context->capset.cap.inheritable = new->cap_effective;
context->capset.cap.permitted = new->cap_permitted;
audit: add ambient capabilities to CAPSET and BPRM_FCAPS records Capabilities were augmented to include ambient capabilities in v4.3 commit 58319057b784 ("capabilities: ambient capabilities"). Add ambient capabilities to the audit BPRM_FCAPS and CAPSET records. The record contains fields "old_pp", "old_pi", "old_pe", "new_pp", "new_pi", "new_pe" so in keeping with the previous record normalizations, change the "new_*" variants to simply drop the "new_" prefix. A sample of the replaced BPRM_FCAPS record: RAW: type=BPRM_FCAPS msg=audit(1491468034.252:237): fver=2 fp=0000000000200000 fi=0000000000000000 fe=1 old_pp=0000000000000000 old_pi=0000000000000000 old_pe=0000000000000000 old_pa=0000000000000000 pp=0000000000200000 pi=0000000000000000 pe=0000000000200000 pa=0000000000000000 INTERPRET: type=BPRM_FCAPS msg=audit(04/06/2017 04:40:34.252:237): fver=2 fp=sys_admin fi=none fe=chown old_pp=none old_pi=none old_pe=none old_pa=none pp=sys_admin pi=none pe=sys_admin pa=none A sample of the replaced CAPSET record: RAW: type=CAPSET msg=audit(1491469502.371:242): pid=833 cap_pi=0000003fffffffff cap_pp=0000003fffffffff cap_pe=0000003fffffffff cap_pa=0000000000000000 INTERPRET: type=CAPSET msg=audit(04/06/2017 05:05:02.371:242) : pid=833 cap_pi=chown,dac_override,dac_read_search,fowner,fsetid,kill, setgid,setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource,sys_time, sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pp=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pe=chown,dac_override,dac_read_search,fowner,fsetid,kill,setgid, setuid,setpcap,linux_immutable,net_bind_service,net_broadcast, net_admin,net_raw,ipc_lock,ipc_owner,sys_module,sys_rawio,sys_chroot, sys_ptrace,sys_pacct,sys_admin,sys_boot,sys_nice,sys_resource, sys_time,sys_tty_config,mknod,lease,audit_write,audit_control,setfcap, mac_override,mac_admin,syslog,wake_alarm,block_suspend,audit_read cap_pa=none See: https://github.com/linux-audit/audit-kernel/issues/40 Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Serge Hallyn <serge@hallyn.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
2017-04-07 14:17:27 +00:00
context->capset.cap.ambient = new->cap_ambient;
context->type = AUDIT_CAPSET;
}
void __audit_mmap_fd(int fd, int flags)
{
struct audit_context *context = audit_context();
context->mmap.fd = fd;
context->mmap.flags = flags;
context->type = AUDIT_MMAP;
}
audit: add OPENAT2 record to list "how" info Since the openat2(2) syscall uses a struct open_how pointer to communicate its parameters they are not usefully recorded by the audit SYSCALL record's four existing arguments. Add a new audit record type OPENAT2 that reports the parameters in its third argument, struct open_how with fields oflag, mode and resolve. The new record in the context of an event would look like: time->Wed Mar 17 16:28:53 2021 type=PROCTITLE msg=audit(1616012933.531:184): proctitle= 73797363616C6C735F66696C652F6F70656E617432002F746D702F61756469742D 7465737473756974652D737641440066696C652D6F70656E617432 type=PATH msg=audit(1616012933.531:184): item=1 name="file-openat2" inode=29 dev=00:1f mode=0100600 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmp_t:s0 nametype=CREATE cap_fp=0 cap_fi=0 cap_fe=0 cap_fver=0 cap_frootid=0 type=PATH msg=audit(1616012933.531:184): item=0 name="/root/rgb/git/audit-testsuite/tests" inode=25 dev=00:1f mode=040700 ouid=0 ogid=0 rdev=00:00 obj=unconfined_u:object_r:user_tmp_t:s0 nametype=PARENT cap_fp=0 cap_fi=0 cap_fe=0 cap_fver=0 cap_frootid=0 type=CWD msg=audit(1616012933.531:184): cwd="/root/rgb/git/audit-testsuite/tests" type=OPENAT2 msg=audit(1616012933.531:184): oflag=0100302 mode=0600 resolve=0xa type=SYSCALL msg=audit(1616012933.531:184): arch=c000003e syscall=437 success=yes exit=4 a0=3 a1=7ffe315f1c53 a2=7ffe315f1550 a3=18 items=2 ppid=528 pid=540 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0 fsgid=0 tty=ttyS0 ses=1 comm="openat2" exe="/root/rgb/git/audit-testsuite/tests/syscalls_file/openat2" subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 key="testsuite-1616012933-bjAUcEPO" Link: https://lore.kernel.org/r/d23fbb89186754487850367224b060e26f9b7181.1621363275.git.rgb@redhat.com Signed-off-by: Richard Guy Briggs <rgb@redhat.com> Acked-by: Christian Brauner <christian.brauner@ubuntu.com> [PM: tweak subject, wrap example, move AUDIT_OPENAT2 to 1337] Signed-off-by: Paul Moore <paul@paul-moore.com>
2021-05-19 20:00:22 +00:00
void __audit_openat2_how(struct open_how *how)
{
struct audit_context *context = audit_context();
context->openat2.flags = how->flags;
context->openat2.mode = how->mode;
context->openat2.resolve = how->resolve;
context->type = AUDIT_OPENAT2;
}
void __audit_log_kern_module(char *name)
{
struct audit_context *context = audit_context();
context->module.name = kstrdup(name, GFP_KERNEL);
if (!context->module.name)
audit_log_lost("out of memory in __audit_log_kern_module");
context->type = AUDIT_KERN_MODULE;
}
void __audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar)
audit: Record fanotify access control decisions The fanotify interface allows user space daemons to make access control decisions. Under common criteria requirements, we need to optionally record decisions based on policy. This patch adds a bit mask, FAN_AUDIT, that a user space daemon can 'or' into the response decision which will tell the kernel that it made a decision and record it. It would be used something like this in user space code: response.response = FAN_DENY | FAN_AUDIT; write(fd, &response, sizeof(struct fanotify_response)); When the syscall ends, the audit system will record the decision as a AUDIT_FANOTIFY auxiliary record to denote that the reason this event occurred is the result of an access control decision from fanotify rather than DAC or MAC policy. A sample event looks like this: type=PATH msg=audit(1504310584.332:290): item=0 name="./evil-ls" inode=1319561 dev=fc:03 mode=0100755 ouid=1000 ogid=1000 rdev=00:00 obj=unconfined_u:object_r:user_home_t:s0 nametype=NORMAL type=CWD msg=audit(1504310584.332:290): cwd="/home/sgrubb" type=SYSCALL msg=audit(1504310584.332:290): arch=c000003e syscall=2 success=no exit=-1 a0=32cb3fca90 a1=0 a2=43 a3=8 items=1 ppid=901 pid=959 auid=1000 uid=1000 gid=1000 euid=1000 suid=1000 fsuid=1000 egid=1000 sgid=1000 fsgid=1000 tty=pts1 ses=3 comm="bash" exe="/usr/bin/bash" subj=unconfined_u:unconfined_r:unconfined_t: s0-s0:c0.c1023 key=(null) type=FANOTIFY msg=audit(1504310584.332:290): resp=2 Prior to using the audit flag, the developer needs to call fanotify_init or'ing in FAN_ENABLE_AUDIT to ensure that the kernel supports auditing. The calling process must also have the CAP_AUDIT_WRITE capability. Signed-off-by: sgrubb <sgrubb@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-10-03 00:21:39 +00:00
{
/* {subj,obj}_trust values are {0,1,2}: no,yes,unknown */
switch (friar->hdr.type) {
case FAN_RESPONSE_INFO_NONE:
audit_log(audit_context(), GFP_KERNEL, AUDIT_FANOTIFY,
"resp=%u fan_type=%u fan_info=0 subj_trust=2 obj_trust=2",
response, FAN_RESPONSE_INFO_NONE);
break;
case FAN_RESPONSE_INFO_AUDIT_RULE:
audit_log(audit_context(), GFP_KERNEL, AUDIT_FANOTIFY,
"resp=%u fan_type=%u fan_info=%X subj_trust=%u obj_trust=%u",
response, friar->hdr.type, friar->rule_number,
friar->subj_trust, friar->obj_trust);
}
audit: Record fanotify access control decisions The fanotify interface allows user space daemons to make access control decisions. Under common criteria requirements, we need to optionally record decisions based on policy. This patch adds a bit mask, FAN_AUDIT, that a user space daemon can 'or' into the response decision which will tell the kernel that it made a decision and record it. It would be used something like this in user space code: response.response = FAN_DENY | FAN_AUDIT; write(fd, &response, sizeof(struct fanotify_response)); When the syscall ends, the audit system will record the decision as a AUDIT_FANOTIFY auxiliary record to denote that the reason this event occurred is the result of an access control decision from fanotify rather than DAC or MAC policy. A sample event looks like this: type=PATH msg=audit(1504310584.332:290): item=0 name="./evil-ls" inode=1319561 dev=fc:03 mode=0100755 ouid=1000 ogid=1000 rdev=00:00 obj=unconfined_u:object_r:user_home_t:s0 nametype=NORMAL type=CWD msg=audit(1504310584.332:290): cwd="/home/sgrubb" type=SYSCALL msg=audit(1504310584.332:290): arch=c000003e syscall=2 success=no exit=-1 a0=32cb3fca90 a1=0 a2=43 a3=8 items=1 ppid=901 pid=959 auid=1000 uid=1000 gid=1000 euid=1000 suid=1000 fsuid=1000 egid=1000 sgid=1000 fsgid=1000 tty=pts1 ses=3 comm="bash" exe="/usr/bin/bash" subj=unconfined_u:unconfined_r:unconfined_t: s0-s0:c0.c1023 key=(null) type=FANOTIFY msg=audit(1504310584.332:290): resp=2 Prior to using the audit flag, the developer needs to call fanotify_init or'ing in FAN_ENABLE_AUDIT to ensure that the kernel supports auditing. The calling process must also have the CAP_AUDIT_WRITE capability. Signed-off-by: sgrubb <sgrubb@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-10-03 00:21:39 +00:00
}
void __audit_tk_injoffset(struct timespec64 offset)
{
struct audit_context *context = audit_context();
ntp: Audit NTP parameters adjustment Emit an audit record every time selected NTP parameters are modified from userspace (via adjtimex(2) or clock_adjtime(2)). These parameters may be used to indirectly change system clock, and thus their modifications should be audited. Such events will now generate records of type AUDIT_TIME_ADJNTPVAL containing the following fields: - op -- which value was adjusted: - offset -- corresponding to the time_offset variable - freq -- corresponding to the time_freq variable - status -- corresponding to the time_status variable - adjust -- corresponding to the time_adjust variable - tick -- corresponding to the tick_usec variable - tai -- corresponding to the timekeeping's TAI offset - old -- the old value - new -- the new value Example records: type=TIME_ADJNTPVAL msg=audit(1530616044.507:7): op=status old=64 new=8256 type=TIME_ADJNTPVAL msg=audit(1530616044.511:11): op=freq old=0 new=49180377088000 The records of this type will be associated with the corresponding syscall records. An overview of parameter changes that can be done via do_adjtimex() (based on information from Miroslav Lichvar) and whether they are audited: __timekeeping_set_tai_offset() -- sets the offset from the International Atomic Time (AUDITED) NTP variables: time_offset -- can adjust the clock by up to 0.5 seconds per call and also speed it up or slow down by up to about 0.05% (43 seconds per day) (AUDITED) time_freq -- can speed up or slow down by up to about 0.05% (AUDITED) time_status -- can insert/delete leap seconds and it also enables/ disables synchronization of the hardware real-time clock (AUDITED) time_maxerror, time_esterror -- change error estimates used to inform userspace applications (NOT AUDITED) time_constant -- controls the speed of the clock adjustments that are made when time_offset is set (NOT AUDITED) time_adjust -- can temporarily speed up or slow down the clock by up to 0.05% (AUDITED) tick_usec -- a more extreme version of time_freq; can speed up or slow down the clock by up to 10% (AUDITED) Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Richard Guy Briggs <rgb@redhat.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-04-10 09:14:20 +00:00
/* only set type if not already set by NTP */
if (!context->type)
context->type = AUDIT_TIME_INJOFFSET;
memcpy(&context->time.tk_injoffset, &offset, sizeof(offset));
ntp: Audit NTP parameters adjustment Emit an audit record every time selected NTP parameters are modified from userspace (via adjtimex(2) or clock_adjtime(2)). These parameters may be used to indirectly change system clock, and thus their modifications should be audited. Such events will now generate records of type AUDIT_TIME_ADJNTPVAL containing the following fields: - op -- which value was adjusted: - offset -- corresponding to the time_offset variable - freq -- corresponding to the time_freq variable - status -- corresponding to the time_status variable - adjust -- corresponding to the time_adjust variable - tick -- corresponding to the tick_usec variable - tai -- corresponding to the timekeeping's TAI offset - old -- the old value - new -- the new value Example records: type=TIME_ADJNTPVAL msg=audit(1530616044.507:7): op=status old=64 new=8256 type=TIME_ADJNTPVAL msg=audit(1530616044.511:11): op=freq old=0 new=49180377088000 The records of this type will be associated with the corresponding syscall records. An overview of parameter changes that can be done via do_adjtimex() (based on information from Miroslav Lichvar) and whether they are audited: __timekeeping_set_tai_offset() -- sets the offset from the International Atomic Time (AUDITED) NTP variables: time_offset -- can adjust the clock by up to 0.5 seconds per call and also speed it up or slow down by up to about 0.05% (43 seconds per day) (AUDITED) time_freq -- can speed up or slow down by up to about 0.05% (AUDITED) time_status -- can insert/delete leap seconds and it also enables/ disables synchronization of the hardware real-time clock (AUDITED) time_maxerror, time_esterror -- change error estimates used to inform userspace applications (NOT AUDITED) time_constant -- controls the speed of the clock adjustments that are made when time_offset is set (NOT AUDITED) time_adjust -- can temporarily speed up or slow down the clock by up to 0.05% (AUDITED) tick_usec -- a more extreme version of time_freq; can speed up or slow down the clock by up to 10% (AUDITED) Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Richard Guy Briggs <rgb@redhat.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-04-10 09:14:20 +00:00
}
void __audit_ntp_log(const struct audit_ntp_data *ad)
{
struct audit_context *context = audit_context();
int type;
for (type = 0; type < AUDIT_NTP_NVALS; type++)
if (ad->vals[type].newval != ad->vals[type].oldval) {
/* unconditionally set type, overwriting TK */
context->type = AUDIT_TIME_ADJNTPVAL;
memcpy(&context->time.ntp_data, ad, sizeof(*ad));
break;
}
ntp: Audit NTP parameters adjustment Emit an audit record every time selected NTP parameters are modified from userspace (via adjtimex(2) or clock_adjtime(2)). These parameters may be used to indirectly change system clock, and thus their modifications should be audited. Such events will now generate records of type AUDIT_TIME_ADJNTPVAL containing the following fields: - op -- which value was adjusted: - offset -- corresponding to the time_offset variable - freq -- corresponding to the time_freq variable - status -- corresponding to the time_status variable - adjust -- corresponding to the time_adjust variable - tick -- corresponding to the tick_usec variable - tai -- corresponding to the timekeeping's TAI offset - old -- the old value - new -- the new value Example records: type=TIME_ADJNTPVAL msg=audit(1530616044.507:7): op=status old=64 new=8256 type=TIME_ADJNTPVAL msg=audit(1530616044.511:11): op=freq old=0 new=49180377088000 The records of this type will be associated with the corresponding syscall records. An overview of parameter changes that can be done via do_adjtimex() (based on information from Miroslav Lichvar) and whether they are audited: __timekeeping_set_tai_offset() -- sets the offset from the International Atomic Time (AUDITED) NTP variables: time_offset -- can adjust the clock by up to 0.5 seconds per call and also speed it up or slow down by up to about 0.05% (43 seconds per day) (AUDITED) time_freq -- can speed up or slow down by up to about 0.05% (AUDITED) time_status -- can insert/delete leap seconds and it also enables/ disables synchronization of the hardware real-time clock (AUDITED) time_maxerror, time_esterror -- change error estimates used to inform userspace applications (NOT AUDITED) time_constant -- controls the speed of the clock adjustments that are made when time_offset is set (NOT AUDITED) time_adjust -- can temporarily speed up or slow down the clock by up to 0.05% (AUDITED) tick_usec -- a more extreme version of time_freq; can speed up or slow down the clock by up to 10% (AUDITED) Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Richard Guy Briggs <rgb@redhat.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-04-10 09:14:20 +00:00
}
void __audit_log_nfcfg(const char *name, u8 af, unsigned int nentries,
enum audit_nfcfgop op, gfp_t gfp)
{
struct audit_buffer *ab;
char comm[sizeof(current->comm)];
ab = audit_log_start(audit_context(), gfp, AUDIT_NETFILTER_CFG);
if (!ab)
return;
audit_log_format(ab, "table=%s family=%u entries=%u op=%s",
name, af, nentries, audit_nfcfgs[op].s);
audit_log_format(ab, " pid=%u", task_pid_nr(current));
audit_log_task_context(ab); /* subj= */
audit_log_format(ab, " comm=");
audit_log_untrustedstring(ab, get_task_comm(comm, current));
audit_log_end(ab);
}
EXPORT_SYMBOL_GPL(__audit_log_nfcfg);
static void audit_log_task(struct audit_buffer *ab)
{
kuid_t auid, uid;
kgid_t gid;
unsigned int sessionid;
char comm[sizeof(current->comm)];
auid = audit_get_loginuid(current);
sessionid = audit_get_sessionid(current);
current_uid_gid(&uid, &gid);
audit_log_format(ab, "auid=%u uid=%u gid=%u ses=%u",
from_kuid(&init_user_ns, auid),
from_kuid(&init_user_ns, uid),
from_kgid(&init_user_ns, gid),
sessionid);
audit_log_task_context(ab);
audit_log_format(ab, " pid=%d comm=", task_tgid_nr(current));
audit_log_untrustedstring(ab, get_task_comm(comm, current));
audit_log_d_path_exe(ab, current->mm);
}
/**
* audit_core_dumps - record information about processes that end abnormally
* @signr: signal value
*
* If a process ends with a core dump, something fishy is going on and we
* should record the event for investigation.
*/
void audit_core_dumps(long signr)
{
struct audit_buffer *ab;
if (!audit_enabled)
return;
if (signr == SIGQUIT) /* don't care for those */
return;
ab = audit_log_start(audit_context(), GFP_KERNEL, AUDIT_ANOM_ABEND);
if (unlikely(!ab))
return;
audit_log_task(ab);
audit_log_format(ab, " sig=%ld res=1", signr);
audit_log_end(ab);
}
/**
* audit_seccomp - record information about a seccomp action
* @syscall: syscall number
* @signr: signal value
* @code: the seccomp action
*
* Record the information associated with a seccomp action. Event filtering for
* seccomp actions that are not to be logged is done in seccomp_log().
* Therefore, this function forces auditing independent of the audit_enabled
* and dummy context state because seccomp actions should be logged even when
* audit is not in use.
*/
void audit_seccomp(unsigned long syscall, long signr, int code)
{
struct audit_buffer *ab;
ab = audit_log_start(audit_context(), GFP_KERNEL, AUDIT_SECCOMP);
if (unlikely(!ab))
return;
audit_log_task(ab);
audit_log_format(ab, " sig=%ld arch=%x syscall=%ld compat=%d ip=0x%lx code=0x%x",
syscall_get_arch: add "struct task_struct *" argument This argument is required to extend the generic ptrace API with PTRACE_GET_SYSCALL_INFO request: syscall_get_arch() is going to be called from ptrace_request() along with syscall_get_nr(), syscall_get_arguments(), syscall_get_error(), and syscall_get_return_value() functions with a tracee as their argument. The primary intent is that the triple (audit_arch, syscall_nr, arg1..arg6) should describe what system call is being called and what its arguments are. Reverts: 5e937a9ae913 ("syscall_get_arch: remove useless function arguments") Reverts: 1002d94d3076 ("syscall.h: fix doc text for syscall_get_arch()") Reviewed-by: Andy Lutomirski <luto@kernel.org> # for x86 Reviewed-by: Palmer Dabbelt <palmer@sifive.com> Acked-by: Paul Moore <paul@paul-moore.com> Acked-by: Paul Burton <paul.burton@mips.com> # MIPS parts Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc) Acked-by: Kees Cook <keescook@chromium.org> # seccomp parts Acked-by: Mark Salter <msalter@redhat.com> # for the c6x bit Cc: Elvira Khabirova <lineprinter@altlinux.org> Cc: Eugene Syromyatnikov <esyr@redhat.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: x86@kernel.org Cc: linux-alpha@vger.kernel.org Cc: linux-snps-arc@lists.infradead.org Cc: linux-arm-kernel@lists.infradead.org Cc: linux-c6x-dev@linux-c6x.org Cc: uclinux-h8-devel@lists.sourceforge.jp Cc: linux-hexagon@vger.kernel.org Cc: linux-ia64@vger.kernel.org Cc: linux-m68k@lists.linux-m68k.org Cc: linux-mips@vger.kernel.org Cc: nios2-dev@lists.rocketboards.org Cc: openrisc@lists.librecores.org Cc: linux-parisc@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Cc: linux-riscv@lists.infradead.org Cc: linux-s390@vger.kernel.org Cc: linux-sh@vger.kernel.org Cc: sparclinux@vger.kernel.org Cc: linux-um@lists.infradead.org Cc: linux-xtensa@linux-xtensa.org Cc: linux-arch@vger.kernel.org Cc: linux-audit@redhat.com Signed-off-by: Dmitry V. Levin <ldv@altlinux.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2019-03-17 23:30:18 +00:00
signr, syscall_get_arch(current), syscall,
in_compat_syscall(), KSTK_EIP(current), code);
audit_log_end(ab);
}
seccomp: Audit attempts to modify the actions_logged sysctl The decision to log a seccomp action will always be subject to the value of the kernel.seccomp.actions_logged sysctl, even for processes that are being inspected via the audit subsystem, in an upcoming patch. Therefore, we need to emit an audit record on attempts at writing to the actions_logged sysctl when auditing is enabled. This patch updates the write handler for the actions_logged sysctl to emit an audit record on attempts to write to the sysctl. Successful writes to the sysctl will result in a record that includes a normalized list of logged actions in the "actions" field and a "res" field equal to 1. Unsuccessful writes to the sysctl will result in a record that doesn't include the "actions" field and has a "res" field equal to 0. Not all unsuccessful writes to the sysctl are audited. For example, an audit record will not be emitted if an unprivileged process attempts to open the sysctl file for reading since that access control check is not part of the sysctl's write handler. Below are some example audit records when writing various strings to the actions_logged sysctl. Writing "not-a-real-action", when the kernel.seccomp.actions_logged sysctl previously was "kill_process kill_thread trap errno trace log", emits this audit record: type=CONFIG_CHANGE msg=audit(1525392371.454:120): op=seccomp-logging actions=? old-actions=kill_process,kill_thread,trap,errno,trace,log res=0 If you then write "kill_process kill_thread errno trace log", this audit record is emitted: type=CONFIG_CHANGE msg=audit(1525392401.645:126): op=seccomp-logging actions=kill_process,kill_thread,errno,trace,log old-actions=kill_process,kill_thread,trap,errno,trace,log res=1 If you then write "log log errno trace kill_process kill_thread", which is unordered and contains the log action twice, it results in the same actions value as the previous record: type=CONFIG_CHANGE msg=audit(1525392436.354:132): op=seccomp-logging actions=kill_process,kill_thread,errno,trace,log old-actions=kill_process,kill_thread,errno,trace,log res=1 If you then write an empty string to the sysctl, this audit record is emitted: type=CONFIG_CHANGE msg=audit(1525392494.413:138): op=seccomp-logging actions=(none) old-actions=kill_process,kill_thread,errno,trace,log res=1 No audit records are generated when reading the actions_logged sysctl. Suggested-by: Steve Grubb <sgrubb@redhat.com> Signed-off-by: Tyler Hicks <tyhicks@canonical.com> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2018-05-04 01:08:14 +00:00
void audit_seccomp_actions_logged(const char *names, const char *old_names,
int res)
{
struct audit_buffer *ab;
if (!audit_enabled)
return;
ab = audit_log_start(audit_context(), GFP_KERNEL,
seccomp: Audit attempts to modify the actions_logged sysctl The decision to log a seccomp action will always be subject to the value of the kernel.seccomp.actions_logged sysctl, even for processes that are being inspected via the audit subsystem, in an upcoming patch. Therefore, we need to emit an audit record on attempts at writing to the actions_logged sysctl when auditing is enabled. This patch updates the write handler for the actions_logged sysctl to emit an audit record on attempts to write to the sysctl. Successful writes to the sysctl will result in a record that includes a normalized list of logged actions in the "actions" field and a "res" field equal to 1. Unsuccessful writes to the sysctl will result in a record that doesn't include the "actions" field and has a "res" field equal to 0. Not all unsuccessful writes to the sysctl are audited. For example, an audit record will not be emitted if an unprivileged process attempts to open the sysctl file for reading since that access control check is not part of the sysctl's write handler. Below are some example audit records when writing various strings to the actions_logged sysctl. Writing "not-a-real-action", when the kernel.seccomp.actions_logged sysctl previously was "kill_process kill_thread trap errno trace log", emits this audit record: type=CONFIG_CHANGE msg=audit(1525392371.454:120): op=seccomp-logging actions=? old-actions=kill_process,kill_thread,trap,errno,trace,log res=0 If you then write "kill_process kill_thread errno trace log", this audit record is emitted: type=CONFIG_CHANGE msg=audit(1525392401.645:126): op=seccomp-logging actions=kill_process,kill_thread,errno,trace,log old-actions=kill_process,kill_thread,trap,errno,trace,log res=1 If you then write "log log errno trace kill_process kill_thread", which is unordered and contains the log action twice, it results in the same actions value as the previous record: type=CONFIG_CHANGE msg=audit(1525392436.354:132): op=seccomp-logging actions=kill_process,kill_thread,errno,trace,log old-actions=kill_process,kill_thread,errno,trace,log res=1 If you then write an empty string to the sysctl, this audit record is emitted: type=CONFIG_CHANGE msg=audit(1525392494.413:138): op=seccomp-logging actions=(none) old-actions=kill_process,kill_thread,errno,trace,log res=1 No audit records are generated when reading the actions_logged sysctl. Suggested-by: Steve Grubb <sgrubb@redhat.com> Signed-off-by: Tyler Hicks <tyhicks@canonical.com> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2018-05-04 01:08:14 +00:00
AUDIT_CONFIG_CHANGE);
if (unlikely(!ab))
return;
audit_log_format(ab,
"op=seccomp-logging actions=%s old-actions=%s res=%d",
names, old_names, res);
seccomp: Audit attempts to modify the actions_logged sysctl The decision to log a seccomp action will always be subject to the value of the kernel.seccomp.actions_logged sysctl, even for processes that are being inspected via the audit subsystem, in an upcoming patch. Therefore, we need to emit an audit record on attempts at writing to the actions_logged sysctl when auditing is enabled. This patch updates the write handler for the actions_logged sysctl to emit an audit record on attempts to write to the sysctl. Successful writes to the sysctl will result in a record that includes a normalized list of logged actions in the "actions" field and a "res" field equal to 1. Unsuccessful writes to the sysctl will result in a record that doesn't include the "actions" field and has a "res" field equal to 0. Not all unsuccessful writes to the sysctl are audited. For example, an audit record will not be emitted if an unprivileged process attempts to open the sysctl file for reading since that access control check is not part of the sysctl's write handler. Below are some example audit records when writing various strings to the actions_logged sysctl. Writing "not-a-real-action", when the kernel.seccomp.actions_logged sysctl previously was "kill_process kill_thread trap errno trace log", emits this audit record: type=CONFIG_CHANGE msg=audit(1525392371.454:120): op=seccomp-logging actions=? old-actions=kill_process,kill_thread,trap,errno,trace,log res=0 If you then write "kill_process kill_thread errno trace log", this audit record is emitted: type=CONFIG_CHANGE msg=audit(1525392401.645:126): op=seccomp-logging actions=kill_process,kill_thread,errno,trace,log old-actions=kill_process,kill_thread,trap,errno,trace,log res=1 If you then write "log log errno trace kill_process kill_thread", which is unordered and contains the log action twice, it results in the same actions value as the previous record: type=CONFIG_CHANGE msg=audit(1525392436.354:132): op=seccomp-logging actions=kill_process,kill_thread,errno,trace,log old-actions=kill_process,kill_thread,errno,trace,log res=1 If you then write an empty string to the sysctl, this audit record is emitted: type=CONFIG_CHANGE msg=audit(1525392494.413:138): op=seccomp-logging actions=(none) old-actions=kill_process,kill_thread,errno,trace,log res=1 No audit records are generated when reading the actions_logged sysctl. Suggested-by: Steve Grubb <sgrubb@redhat.com> Signed-off-by: Tyler Hicks <tyhicks@canonical.com> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Paul Moore <paul@paul-moore.com>
2018-05-04 01:08:14 +00:00
audit_log_end(ab);
}
struct list_head *audit_killed_trees(void)
{
struct audit_context *ctx = audit_context();
if (likely(!ctx || ctx->context == AUDIT_CTX_UNUSED))
return NULL;
return &ctx->killed_trees;
}