Recursive chown/chmod not showing enclosing directory
by Mark Moseley
Hi. Didn't realize this was a closed list till I started wondering why
the automated invite hadn't shown up :)
Just a quick question: I'm working on parsing audit logs and I noticed
one oddity. If a chown or chmod is done recursively on a directory
(and I imagine there are other examples than chown/chmod) done from
somewhere outside the affected directory, in the audit log entries for
the *files* within those directories, there's no way to track back
what directory the files live in.
Example:
>From /tmp, doing a "chown -R 0 /home/moseley/tmp/tmp/", where the
contents of /home/moseley/tmp/tmp/ are two files, 'a' and 'b' (I've
added spacing to make it more readable):
type=SYSCALL msg=audit(1349375745.138:4143): arch=c000003e syscall=260
success=yes exit=0 a0=4 a1=77a988 a2=0 a3=ffffffff items=1 ppid=5775
pid=19589 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0
fsgid=0 tty=pts13 ses=3387 comm="chown" exe="/bin/chown" key=(null)
type=CWD msg=audit(1349375745.138:4143): cwd="/tmp"
type=PATH msg=audit(1349375745.138:4143): item=0 name="b"
inode=2367514 dev=08:02 mode=0100664 ouid=1000 ogid=1000 rdev=00:00
---
type=SYSCALL msg=audit(1349375745.138:4144): arch=c000003e syscall=260
success=yes exit=0 a0=4 a1=782b08 a2=0 a3=ffffffff items=1 ppid=5775
pid=19589 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0 egid=0 sgid=0
fsgid=0 tty=pts13 ses=3387 comm="chown" exe="/bin/chown" key=(null)
type=CWD msg=audit(1349375745.138:4144): cwd="/tmp"
type=PATH msg=audit(1349375745.138:4144): item=0 name="a"
inode=2367514 dev=08:02 mode=0100664 ouid=0 ogid=1000 rdev=00:00
---
type=SYSCALL msg=audit(1349375745.138:4145): arch=c000003e syscall=260
success=yes exit=0 a0=ffffffffffffff9c a1=779520 a2=0 a3=ffffffff
items=1 ppid=5775 pid=19589 auid=0 uid=0 gid=0 euid=0 suid=0 fsuid=0
egid=0 sgid=0 fsgid=0 tty=pts13 ses=3387 comm="chown" exe="/bin/chown"
key=(null)
type=CWD msg=audit(1349375745.138:4145): cwd="/tmp"
type=PATH msg=audit(1349375745.138:4145): item=0
name="/home/moseley/tmp/tmp/" inode=2367031 dev=08:02 mode=040775
ouid=1000 ogid=1000 rdev=00:00
In the first two entries, there's no indication that 'a' or 'b' live
in /home/moseley/tmp/tmp/.
In the case of a non-relative chown/chmod (e.g. chown'ing
/home/moseley/tmp/tmp/a), the 'item=0' line has the full pathname to
/home/moseley/tmp/tmp/a in its entry.
My question is, is there anything in these entries that I might be
missing that ties the first two back to the 3rd? I'm building a
tracking system for 'changed/written' files, so I change all 'item'
pathnames into absolute pathnames, but in this case, I'm at a loss as
to how to (programmatically) tie them together, beyond "pid". Or
better yet, is there some flag in the first two entries that I might
be missing that shows that they're 'children' of the third entry.
I'm not crazy about the idea of building some of state into my parsing
(i.e. I see a relative chown/chmod 'item' entry, so stash these and
keep watching for chown/chmod's with the same PID) but if that's my
only option, then that's my only option. But I figured I'd ask to see
if there was something far more clever than I'm missing (possibly in
plain sight).
Thanks!
12 years
- 2
- 2
[PATCH v7 00/49] audit/getname/estale patch series
by Jeff Layton
This patchset is a unification of the 3 series that I posted
individually over the last several months. The main reason for all of
this is the last series (the estale one). The 3 series are as follows:
audit: this is an overhaul of the audit_names code that cleans up the
code, adds a few micro-optimizations and fixes a few minor bugs. It also
changes that code to allow for retrying syscalls without creating
duplicate records.
getname: this changes the getname/putname interface to deal with a new
struct getname_info rather than "bare" strings. This allows us to pass
around some ancillary data with the resulting getname string. It also
has the nice effect of making it OK to call getname on the same userland
string more than once.
estale: this patchset retrofits many of the path-based syscalls in the
kernel to retry the lookup and operation when the operation returns
ESTALE. There might be a few more that need similar changes afterward,
but this should cover most of the ones people are interested in.
Because there are dependencies between these patchsets, I decided to
just post the whole shebang as one giant patchset.
Al's recent execve unification work paved the way for this by greatly
reducing the number of getname() callers. Thus, this set depends on his
work being merged.
At this point, I think it's this set is ready for merge into 3.7 with
one caveat: The getname patches touch some arch specific code in execve
implementations that have not been unified yet. Some of them I don't
have hardware or working emulators for, so they're not even compile
tested.
That said, the arch-specific changes are pretty straightforward. I don't
expect problems with them, but if there are I'm fairly certain we can
straighten them out before 3.7 ships.
Since this is a large set, it may be easier to just pull them. Pull
request follows:
------------------------[snip]---------------------------
The following changes since commit 659c04881db5f69c8f6f789106be4af85404f03b:
sparc32: switch to generic sys_execve() (2012-10-01 01:02:23 -0400)
are available in the git repository at:
git://git.samba.org/jlayton/linux.git estale
for you to fetch changes up to e7c07a567a6ba18ec7d1ef8ceea4f9978a325bed:
vfs: make lremovexattr retry once on ESTALE error (2012-10-01 19:59:58 -0400)
------------------------[snip]---------------------------
Eric Paris (1):
audit: make audit_compare_dname_path use parent_len helper
Jeff Layton (48):
audit: remove unnecessary NULL ptr checks from do_path_lookup
audit: pass in dentry to audit_copy_inode wherever possible
audit: no need to walk list in audit_inode if name is NULL
audit: reverse arguments to audit_inode_child
audit: add a new "type" field to audit_names struct
audit: set the name_len in audit_inode for parent lookups
audit: remove dirlen argument to audit_compare_dname_path
audit: optimize audit_compare_dname_path
audit: overhaul __audit_inode_child to accomodate retrying
vfs: allocate page instead of names_cache buffer in mount_block_root
vfs: make dir_name arg to do_mount a const char *
acct: constify the name arg to acct_on
vfs: define getname_info struct and have getname() return it
audit: allow audit code to satisfy getname requests from its
names_list
vfs: turn do_path_lookup into wrapper around getname_info variant
vfs: make path_openat take a getname_info pointer
audit: make audit_inode take getname_info
vfs: embed getname_info inside of names_cache allocation if possible
vfs: unexport getname and putname symbols
vfs: add a retry_estale helper function to handle retries on ESTALE
vfs: make fstatat retry on ESTALE errors from getattr call
vfs: fix readlinkat to retry on ESTALE
vfs: add new "reval" argument to kern_path_create and
user_path_create
vfs: fix mknodat to retry on ESTALE errors
vfs: fix mkdir to retry on ESTALE errors
vfs: fix symlinkat to retry on ESTALE errors
vfs: fix linkat to retry on ESTALE errors
vfs: add a reval argument to user_path_parent
vfs: make rmdir retry on ESTALE errors
vfs: make do_unlinkat retry on ESTALE errors
vfs: fix renameat to retry on ESTALE errors
vfs: have do_sys_truncate retry once on an ESTALE error
vfs: have faccessat retry once on an ESTALE error
vfs: have chdir retry lookup and call once on ESTALE error
vfs: make chroot retry once on ESTALE error
vfs: make fchmodat retry once on ESTALE errors
vfs: make fchownat retry once on ESTALE errors
vfs: convert do_filp_open to use retry_estale helper
vfs: convert do_file_open_root to use retry_estale helper
vfs: allow utimensat() calls to retry once on an ESTALE error
vfs: allow setxattr to retry once on ESTALE errors
vfs: allow lsetxattr() to retry once on ESTALE errors
vfs: make getxattr retry once on an ESTALE error
vfs: make lgetxattr retry once on ESTALE
vfs: make listxattr retry once on ESTALE error
vfs: make llistxattr retry once on ESTALE error
vfs: make removexattr retry once on ESTALE
vfs: make lremovexattr retry once on ESTALE error
arch/alpha/kernel/osf_sys.c | 16 +-
arch/avr32/kernel/process.c | 4 +-
arch/blackfin/kernel/process.c | 4 +-
arch/cris/arch-v10/kernel/process.c | 4 +-
arch/cris/arch-v32/kernel/process.c | 4 +-
arch/h8300/kernel/process.c | 4 +-
arch/hexagon/kernel/syscall.c | 4 +-
arch/ia64/kernel/process.c | 4 +-
arch/m32r/kernel/process.c | 4 +-
arch/microblaze/kernel/sys_microblaze.c | 4 +-
arch/mips/kernel/linux32.c | 4 +-
arch/mips/kernel/syscall.c | 4 +-
arch/openrisc/kernel/process.c | 4 +-
arch/parisc/hpux/fs.c | 4 +-
arch/parisc/kernel/process.c | 4 +-
arch/parisc/kernel/sys_parisc32.c | 4 +-
arch/powerpc/platforms/cell/spufs/syscalls.c | 2 +-
arch/score/kernel/sys_score.c | 4 +-
arch/sh/kernel/process_32.c | 4 +-
arch/sh/kernel/process_64.c | 4 +-
arch/tile/kernel/process.c | 8 +-
arch/unicore32/kernel/sys.c | 4 +-
arch/xtensa/kernel/process.c | 4 +-
drivers/base/devtmpfs.c | 7 +-
fs/btrfs/ioctl.c | 2 +-
fs/compat.c | 12 +-
fs/exec.c | 14 +-
fs/filesystems.c | 4 +-
fs/internal.h | 4 +-
fs/namei.c | 469 +++++++++++++++++----------
fs/namespace.c | 6 +-
fs/ocfs2/refcounttree.c | 3 +-
fs/open.c | 289 ++++++++++-------
fs/quota/quota.c | 4 +-
fs/stat.c | 32 +-
fs/utimes.c | 15 +-
fs/xattr.c | 160 +++++----
include/linux/audit.h | 52 ++-
include/linux/fs.h | 44 ++-
include/linux/fsnotify.h | 8 +-
include/linux/namei.h | 4 +-
init/do_mounts.c | 7 +-
ipc/mqueue.c | 17 +-
kernel/acct.c | 6 +-
kernel/audit.h | 7 +-
kernel/audit_watch.c | 3 +-
kernel/auditfilter.c | 65 ++--
kernel/auditsc.c | 233 ++++++++-----
mm/swapfile.c | 11 +-
net/unix/af_unix.c | 2 +-
50 files changed, 988 insertions(+), 598 deletions(-)
--
1.7.11.4
12 years, 1 month
- 2
- 55