オープンソース・ソフトウェアの開発とダウンロード

Subversion リポジトリの参照

Diff of /trunk/1.8.x/ccs-patch/README.ccs

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

trunk/1.6.x/ccs-patch/README.ccs revision 2562 by kumaneko, Tue May 19 01:11:55 2009 UTC trunk/1.7.x/ccs-patch/README.ccs revision 3131 by kumaneko, Tue Nov 3 03:51:07 2009 UTC
# Line 16  and you enjoy the life with Linux. Line 16  and you enjoy the life with Linux.
16    
17  This project was very inspired by the comic "Card Captor SAKURA",  This project was very inspired by the comic "Card Captor SAKURA",
18  one of the CLAMP's masterworks.  one of the CLAMP's masterworks.
 The names SAKURA and TOMOYO and SYAORAN were borrowed from the comic  
 with the heartfelt thanks to CLAMP.  
   
   
19    
20  ChangeLog:  ChangeLog:
21    
# Line 197  Fix 2006/02/27 Line 193  Fix 2006/02/27
193      @ Fix handling of trailing '\*' in PathMatchesToPattern().      @ Fix handling of trailing '\*' in PathMatchesToPattern().
194    
195        PathMatchesToPattern("/tmp/", "/tmp/\*") returned true        PathMatchesToPattern("/tmp/", "/tmp/\*") returned true
196        because "\*" matches to "more than or equals to 0 character        because "\*" matches "zero or more repetitions of characters
197        until '/' or end". But since this is a comparison between        until '/' or end". But since this is a comparison between
198        directory and non-directory, this should not match.        directory and non-directory, this should not match.
199    
# Line 408  Fix 2006/07/14 Line 404  Fix 2006/07/14
404    
405      @ Change behavior of pathname pattern matching.      @ Change behavior of pathname pattern matching.
406    
407        Until now, it was impossible to use patterns like "\*.txt"        Until now, it was impossible to use patterns like "\*.txt" because
408        because "\*" matched to more than 0 characters until next '/'.        "\*" matched zero or more repetitions of characters until next '/'.
409        Now, "\*" matches to more than 0 characters.        Now, "\*" matches zero or more repetitions of characters.
410    
411        Until now, it was impossible to use patterns like "\$00"        Until now, it was impossible to use patterns like "\$00"
412        because "\$" matched to more than 1 digits until next        because "\$" matched one or more repetitions of digits until next
413        non digit character.        non digit character.
414        Now, "\$" matches to more than 1 digits.        Now, "\$" matches one or more repetitions of digits.
415    
416        Also, new patterns "\x" "\X" "\a" "\A" "\@" are added.        Also, new patterns "\x" "\X" "\a" "\A" "\@" are added.
417    
# Line 432  Fix 2006/07/25 Line 428  Fix 2006/07/25
428    
429      @ Change matching rule for CheckFileACL().      @ Change matching rule for CheckFileACL().
430    
431        Until now, only first entry that matched to the requested pathname        Until now, only first entry that matched the requested pathname
432        was used for permission checking. For example, two entries        was used for permission checking. For example, two entries
433    
434        "2 /tmp/file-\$.txt"        "2 /tmp/file-\$.txt"
# Line 448  Fix 2006/07/25 Line 444  Fix 2006/07/25
444        This may potentially cause trouble because the result of        This may potentially cause trouble because the result of
445        permission checks depends on the order of entries.        permission checks depends on the order of entries.
446    
447        Now, all entries that matched to the requested pathname        Now, all entries that matched the requested pathname
448        are used for permission checking so that the result of        are used for permission checking so that the result of
449        permission checks doesn't depend on the order of entries.        permission checks doesn't depend on the order of entries.
450    
# Line 496  Fix 2006/08/10 Line 492  Fix 2006/08/10
492          "6 /home/\*/\* if task.uid=path1.uid"          "6 /home/\*/\* if task.uid=path1.uid"
493    
494        allows the domain to read-write user's home directory        allows the domain to read-write user's home directory
495        only when the file's owner matches to the process's uid.        only when the file's owner matches the process's uid.
496    
497  Fix 2006/08/22  Fix 2006/08/22
498    
# Line 762  Fix 2006/12/22 Line 758  Fix 2006/12/22
758        if /some/program is called from "some_domain" domain.        if /some/program is called from "some_domain" domain.
759    
760        "some_domain" can be just the last component of domainname.        "some_domain" can be just the last component of domainname.
761        For example, giving "/bin/mail" as "some_domain" matches to        For example, giving "/bin/mail" as "some_domain" matches
762        all domains whose domainname ends with "/bin/mail".        all domains whose domainname ends with "/bin/mail".
763    
764  Fix 2007/01/19  Fix 2007/01/19
# Line 1922  Fix 2009/04/07 Line 1918  Fix 2009/04/07
1918    
1919        This problem happens on little endian platforms (e.g. x86).        This problem happens on little endian platforms (e.g. x86).
1920    
 Fix 2009/04/20  
   
     @ Update recvmsg() hooks.  
   
       Since 1.5.0, I was doing network access control for incoming UDP and RAW  
       packets inside skb_recv_datagram(). But to synchronize with LSM version,  
       I moved ccs_recv_datagram_permission() hook from skb_recv_datagram() to  
       udp_recvmsg()/udpv6_recvmsg()/raw_recvmsg()/rawv6_recvmsg() with name  
       change to ccs_recvmsg_permission().  
   
1921  Fix 2009/05/08  Fix 2009/05/08
1922    
1923      @ Add condition for symlink's target pathname.      @ Add condition for symlink's target pathname.
# Line 1977  Fix 2009/05/19 Line 1963  Fix 2009/05/19
1963    
1964        I modified to unlock the mutex before calling get_fs_type().        I modified to unlock the mutex before calling get_fs_type().
1965    
1966  Version 1.6.8 2009/05/??   Feature enhancement release.  Fix 2009/05/20
1967    
1968        @ Update recvmsg() hooks.
1969    
1970          Since 1.5.0, I was doing network access control for incoming UDP and RAW
1971          packets inside skb_recv_datagram(). But to synchronize with LSM version,
1972          I moved ccs_recv_datagram_permission() hook from skb_recv_datagram() to
1973          udp_recvmsg()/udpv6_recvmsg()/raw_recvmsg()/rawv6_recvmsg() with name
1974          change to ccs_recvmsg_permission().
1975    
1976    Version 1.6.8 2009/05/28   Feature enhancement release.
1977    
1978    Fix 2009/07/03
1979    
1980        @ Fix buffer overrun when used with CONFIG_SLOB=y .
1981    
1982          Since 1.6.7 , ccs_allocate_execve_entry() was requesting for only 4000
1983          bytes while the comment says it is 4096 bytes. This may lead to buffer
1984          overrun when slob allocator is used, for slob allocator allocates exactly
1985          4000 bytes whereas slab and slub allocators allocate 4096 bytes.
1986    
1987    Fix 2009/09/01
1988    
1989        @ Add garbage collector support.
1990    
1991          Until now, it was impossible to release memory used by deleted policy.
1992          I added SRCU based garbage collector so that memory used by deleted
1993          policy will be automatically released.
1994    
1995        @ Remove word length limitation and line length limitation.
1996    
1997          Until now, the max length of a word is 4000 and the max length of a line
1998          is 8192. To be able to handle longer pathnames, I removed these
1999          limitations. Now, the max length (except the domainname and
2000          argv[]/envp[]) is 128K (which is the max amount of memory kmalloc()
2001          can allocate in most environments).
2002    
2003        @ Support more fine grained profile configuration.
2004    
2005          Profile was reconstructed.
2006    
2007        @ Support more fine grained parameters restrictions.
2008    
2009          "allow_create", "allow_mkdir", "allow_mkfifo", "allow_mksock" check
2010          create mode. "allow_mkblock" and "allow_mkchar" check create mode and
2011          major/minor device numbers. "allow_chmod" check new mode. "allow_chown"
2012          checks new owner. "allow_chgrp" checks new group.
2013    
2014        @ Allow number grouping.
2015    
2016          To help specifying numeric values, a new directive "number_group" is
2017          introduced.
2018    
2019        @ Remove "alias" directive and "allow_argv0" directive.
2020    
2021          Until now, "allow_execute" used dereferenced pathname if it is a symlink
2022          unless explicitly specified by "alias" directive.
2023    
2024          Now, "allow_execute" uses symlink's pathname if it is a symlink.
2025          "exec.realpath" in "if" clause checks the dereferenced pathname.
2026          "exec.argv[0]" in "if" clause checks the invocation name.
2027    
2028        @ Remove /proc/ccs/system_policy and /etc/ccs/system_policy.conf .
2029    
2030          "deny_autobind" was moved to /proc/ccs/exception_policy and
2031          /etc/ccs/exception_policy.conf . Other directives were moved to
2032          /proc/ccs/domain_policy and /etc/ccs/domain_policy.conf .
2033    
2034        @ Remove syaoran filesystem.
2035    
2036          Since "allow_create"/"allow_mkdir"/"allow_mkfifo"/"allow_mksock"/
2037          "allow_mkblock"/"allow_mkchar"/"allow_chmod"/"allow_chown"/"allow_chgrp"
2038          can restrict mode changes and owner/group changes, there is no need to
2039          restrict these changes at filesystem level.
2040    
2041          Thus, I removed syaoran filesystem.
2042    
2043        @ Reduce spinlocks.
2044    
2045          Until now, TOMOYO was using own list for detecting memory leak. But as
2046          kernel 2.6.31 introduced memory leak detection mechanism
2047          ( CONFIG_DEBUG_KMEMLEAK ), TOMOYO no longer needs to use own list.
2048    
2049          I removed the list to ruduce use of spinlocks.
2050    
2051        @ Rewrite ccs-patch-2.\*.diff .
2052    
2053          ccs-patch-2.\*.diff was rewriteen like LSM hooks.
2054    
2055        @ Don't check "allow_read/write" for open-for-ioctl-only.
2056    
2057          open(pathname, 3) means open for ioctl() only.
2058          Until now, TOMOYO was checking "allow_read/write" for open(pathname, 3).
2059          But since TOMOYO checks "allow_ioctl" for ioctl(), I modified not to
2060          require "allow_read/write" for open(pathname, 3).
2061    
2062        @ Add missing sigqueue() and tgsigqueue() hooks.
2063    
2064          Until now, kill(), tkill(), tgkill() had hooks but sigqueue() and
2065          tgsigqueue() didn't.
2066    
2067        @ Move files from fs/ to security/ccsecurity.
2068    
2069          Config menu section changed from "File systems" to "Security options".
2070    
2071          Kernel config symbols changed from CONFIG_SAKURA CONFIG_TOMOYO
2072          CONFIG_SYAORAN to CONFIG_CCSECURITY .
2073    
2074        @ Add global PID to audit logs.
2075    
2076          ccs-queryd was using domainname for reaching the domain which the process
2077          belongs to, but the domain could be deleted while ccs-queryd is handling
2078          policy violation. If the domain is deleted, ccs-queryd no longer can
2079          reach the domain by domainname. Thus, ccs-queryd now uses PID for
2080          reaching the domain which the process belongs to.
2081    
2082          Kernel 2.6.24 introduced PID namespace. The PID in access logs generated
2083          by a process inside a container is useless for ccs-queryd for reaching
2084          the domain which the process belongs to.
2085    
2086          Thus, I added global PID in audit logs.
2087          
2088        @ Transit to new domain before do_execve() succeeds.
2089    
2090          Permission checks for interpreters and environment variables are
2091          done using new domain. In order to be allow ccs-queryd to reach the new
2092          domain via global PID, I reverted "Don't transit to new domain until
2093          do_execve() succeeds." made on 2008/10/07.
2094    
2095    Version 1.7.0 2009/09/03   Feature enhancement release.
2096    
2097    Fix 2009/09/04
2098    
2099        @ Fix wrong ccs_profile() calls.
2100    
2101          I can't call ccs_profile() for profile existence test because
2102          ccs_profile() never returns NULL.
2103    
2104    Fix 2009/09/06
2105    
2106        @ Fix wrong error code in ccs_try_alt_exec().
2107    
2108          ccs_try_alt_exec() was returning ENOMEM when kmalloc() failed.
2109          It needs to return -ENOMEM to fail.
2110    
2111    Fix 2009/09/10
2112    
2113        @ Do not check umount() permission for mount(MS_MOVE) requests.
2114    
2115          Until 1.6.x , umount() restriction was black listing. In 1.7.0 , it is
2116          white listing. This change caused "mount --move old new" requests to
2117          require "allow_unmount old" permission in addition to
2118          "allow_mount old new --move 0" permission.
2119          But we don't want to allow umount(old) requests when we want to allow
2120          only mount(old, new, MS_MOVE) requests. Thus, I modified not to check
2121          "allow_unmount old" permission for mount(old, new, MS_MOVE) requests.
2122    
2123    Fix 2009/09/11
2124    
2125        @ Support recursive match operators.
2126    
2127          Until now, ccs_path_matches_pattern() did not support recursive
2128          comparison. Thus, users had to repeat "/\*" when they want to specify
2129          recursively.
2130    
2131          I introduced "\{" and "\}" as repetition operator.
2132          To ensure consistency with TOMOYO's '/'-tokenized pattern matching rules
2133          and "\-" operator, only "/\{dir\}/" sequences (where dir does not contain
2134          '/') is permitted.
2135    
2136    Fix 2009/09/24
2137    
2138        @ Don't check chmod/chown capability for requests from kernel.
2139    
2140          Until now, ccs_setattr_permission() was inserted in notify_change().
2141          But notify_change() is also called by requests from kernel (e.g. UnionFS)
2142          and it made difficult to use TOMOYO on UnionFS.
2143    
2144          Thus, I moved ccs_capable() checks from ccs_setattr_permission() to
2145          ccs_chmod_permission() and ccs_chown_permission(), and removed
2146          ccs_setattr_permission().
2147    
2148    Fix 2009/09/25
2149    
2150        @ Embed more information into audit logs.
2151    
2152          Until now, /proc/ccs/grant_log /proc/ccs/reject_log /proc/ccs/query were
2153          not printing file's information (e.g. file's uid/gid/mode).
2154    
2155          Recently, users who started using "if" clause expect that the learning
2156          mode automatically adds various conditions like "if task.uid=path1.uid".
2157    
2158          But the profile will become too complicated if I support all possible
2159          conditions. Thus, I added all information which is enough to generate
2160          "if" clause with all possible conditions from audit logs.
2161    
2162          Now, the learning mode got different usage. Users can specify
2163          "CONFIG::learning={ max_entry=0 }" in the profile. All requests which
2164          are not permitted by policy will be sent to /proc/ccs/reject_log with
2165          "mode=learning" header lines. Users can selectively append conditions
2166          and append to the policy using "/usr/sbin/ccs-loadpolicy -d".
2167          The learning mode with "CONFIG::learning={ max_entry=0 }" is almost
2168          the same with the permissive mode, only difference is "mode=learning"
2169          and "mode=permissive".
2170    
2171    Fix 2009/10/05
2172    
2173        @ Fix size truncation bug at ccs_memcmp().
2174    
2175          ccs_memcmp() was using "u8" for size parameter by error. Therefore, when
2176          size >= 256 was passed to ccs_memcmp(), it was doing partial comarison
2177          (incorrect result) or read overrun (CPU stall).
2178    
2179          ccs_memcmp() should use "size_t" for size parameter because size of
2180          "struct ccs_condition" may exceed 256 bytes if complicated condition was
2181          given.
2182    
2183    Fix 2009/10/08
2184    
2185        @ Add CONFIG_CCSECURITY_DEFAULT_LOADER option.
2186    
2187          I made the default policy loader's pathname ( /sbin/ccs-init )
2188          configurable.
2189    
2190        @ Add CONFIG_CCSECURITY_ALTERNATIVE_TRIGGER option.
2191    
2192          Some environments do not have /sbin/init . In such environments, we need
2193          to use different program's pathname (e.g. /init or /linuxrc ) as
2194          activation trigger.
2195    
2196          Thus, I made the alternative trigger ( /sbin/ccs-start ) configurable.
2197    
2198    Fix 2009/11/02
2199    
2200        @ Fix buffer contention.
2201    
2202          A permission like
2203    
2204            allow_env PATH if exec.envp["PATH"]="/"
2205    
2206          was not working since I was using the same buffer for both environment
2207          variable's name and value.
2208    
2209    Fix 2009/11/03
2210    
2211        @ Fix memory leak in ccs_write_address_group_policy().
2212    
2213          I forgot to call kfree() if same entry was added.
2214    
2215        @ Reduce mutexes.
2216    
2217          I was using mutex_lock()/mutex_unlock() so that I can use
2218          atomic_dec_and_test() for removing an element from a list.
2219          I moved that operation to garbage collector in order to reduce frequency
2220          of mutex_lock()/mutex_unlock() calls.

Legend:
Removed from v.2562  
changed lines
  Added in v.3131

Back to OSDN">Back to OSDN
ViewVC Help
Powered by ViewVC 1.1.26