Age | Commit message (Collapse) | Author | |
---|---|---|---|
2006-09-19 | Use S_IS* macros insted of masking with S_IF* flags. The latter may | Otto Moerbeek | |
have multiple bits set, which leads to surprising results. Spotted by/partly from Paul Stoeber, more to come. ok ho@ miod@ hshoexer@ | |||
2006-07-02 | sync with systrace 1.6d, keeping local changes | Nikolay Sturm | |
tests and feedback by a few | |||
2006-04-26 | ARGSUSED and a FALLTHROUGH to please lint | Nikolay Sturm | |
ok deraadt | |||
2006-03-06 | convert permanent privilege revocation to use setresuid/setresgid; | Damien Miller | |
ok henning@ | |||
2004-07-07 | fix an issue when scripts are exec'd under systrace where | marius eriksen | |
the argv[0] would be normalized, and hence break scripts that depend on how they were called. this fixes an issue in the ports builds. ok provos@ deraadt@; lots of testing during hackathon sturm@ naddy@ | |||
2004-06-24 | if we have detached after an exec, bail out early, and avoid a double free. | marius eriksen | |
ok provos@, "works fine" sturm@ | |||
2004-06-23 | a few fixes to systrace | marius eriksen | |
- add an exec message so that whenever a set-uid/gid process exec's a new image which we may control, the exec does not go by unnoticed. - take special care to check for P_SUGIDEXEC as well as P_SUGID, corresponding to the same changes that were made in the ptrace code a while ago ok niels@, sturm@; thanks to naddy for testing | |||
2004-03-30 | if a string considered a filename is too long, don't exit but just let | Nikolay Sturm | |
the syscall fail, it might not actually be a filename fixes pr 3140, ok provos@ | |||
2004-01-30 | The empty filename does not receive normalization. | Nikolay Sturm | |
System calls are supposed to fail on it. from provos@, ok markus@ | |||
2003-10-18 | typos from Jared Yanovich; | Jason McIntyre | |
2003-10-08 | originally from cb@openbsd.org, adapted by provos | Nikolay Sturm | |
itojun@ ok fix a race condition between path resolution in userland and the subsequent namei(): inform the kernel portion of valid filenames and then disallow symlink lookups for those filenames by means of a hook in namei(). with suggestions from provos@ also, add (currently unused) seqnr field to struct systrace_replace, from provos@ | |||
2003-08-04 | several diffs from Niels as applied to NetBSD | Nikolay Sturm | |
monkey.org/NetBSD commit messages: - get rid of retarded CWD handling. CWD is fixed to the CWD of the systrace that started everything. - normalize file name function - normalize CWD for cases where CWD has a symlink in it. should solve problems where CWD policies would not match. - avoid warning due to name collision. - fixed contrived race condition during attachment; from marius@monkey.org itojun@ ok | |||
2003-07-19 | - sync with NetBSD or Niels' tarball where appropriate | Nikolay Sturm | |
- keeps local changes - fixes a bug in profile feedback optimization and avoids symbol conflicts with errno - new feature: "ask" action itojun@ ok | |||
2003-06-16 | - limited number of processes per systrace | Jun-ichiro itojun Hagino | |
- escape fixes for special characters markus, sturm ok. from provos | |||
2003-05-17 | pull in a bugfix from systrace-current to let systrace deal with creating | Nikolay Sturm | |
directories correctly OK itojun@, thanks to niels for the help | |||
2003-02-20 | Fix a crash in the systrace found by form@ | Artur Grabowski | |
One is a kernel fix that changes the lockin and one is a userland fix that prevents dereferencing a freed pointer. From provos deraadt@ ok | |||
2002-12-09 | prevent the use of permit for aliases. from provos | Jun-ichiro itojun Hagino | |
2002-11-26 | performance improvement by omitting a redundant getcwd. | Jun-ichiro itojun Hagino | |
from provos | |||
2002-11-12 | fix bug in determining execve name. from provos | Jun-ichiro itojun Hagino | |
2002-10-17 | little cleanup (intercept_getpid dies within the function on error). | Jun-ichiro itojun Hagino | |
from provos | |||
2002-10-16 | support for privilege elevation. | Jun-ichiro itojun Hagino | |
with privilege elevation no suid or sgid binaries are necessary any longer. Applications can be executed completely unprivileged. Systrace raises the privileges for a single system call depending on the configured policy. Idea from discussions with Perry Metzger, Dug Song and Marcus Watts. from provos | |||
2002-10-09 | predicates are part of the grammar now; in non-root case, predicates are | Jun-ichiro itojun Hagino | |
evaluated only once; in root case, predicates and variable expansion are dynamic. from provos | |||
2002-09-17 | daemon should not change the directory. from provos | Jun-ichiro itojun Hagino | |
2002-09-06 | standalone ; at top scope is illegal in ansi c | Theo de Raadt | |
2002-08-28 | fix systrace with chroot. from provos | Jun-ichiro itojun Hagino | |
2002-08-08 | if getcwd fails and we continue dont restcwd. | Niels Provos | |
2002-08-07 | deal better with interrupted system calls | Niels Provos | |
2002-08-05 | uid and gid are not guaranteed to be aligned on a register_t sized boundary. | Jason Wright | |
Use a temporary location and then copy the value into place. provos ok. | |||
2002-08-05 | increase buffer size for getstring, useful for execve arguments. | Niels Provos | |
intercept_filename deals better with symlinked last component lookups. change some translations to use unlinkname. | |||
2002-08-04 | keep track of ppid and allow matching rules to be logged via syslog. | Niels Provos | |
2002-08-02 | performance improvement, reduces number of ioctl calls | Niels Provos | |
2002-08-01 | correctly separate execve argv arguments. increase buffer size for | Niels Provos | |
get_string | |||
2002-08-01 | the last component in a filename for unlink may be a symlink | Niels Provos | |
2002-07-30 | sometimes no-return syscalls (execve) emit errno < 0. ignore them. | Jun-ichiro itojun Hagino | |
2002-07-30 | solve a problem with realpath when the last component of the path is | Niels Provos | |
a directory without S_IXUSR; tested by me and dugsong. | |||
2002-07-22 | add seqnr to message from kernel, userland needs to quote correct seqnr. | Niels Provos | |
avoids problems where tsleep has been interrupted by a signal. | |||
2002-07-19 | constify, have missing prototypes, use pedantic compilation options. | Jun-ichiro itojun Hagino | |
niels ok | |||
2002-07-16 | internal uid/gid tracking. permit can not detach systrace, useful for | Niels Provos | |
sshd. | |||
2002-07-12 | some clean up. install argument replacements only if we are going to | Niels Provos | |
permit the system call. translate some set[e]{g,u}id calls | |||
2002-07-10 | make it work with chroot | Niels Provos | |
2002-07-10 | do not close fds in daemon. | Niels Provos | |
2002-07-09 | allow systrace to run in the background if possible so that the executed | Niels Provos | |
process gets the terminal correctly and exit status reporting works; based on a diff from atatat@atatdot.net from netbsd. | |||
2002-07-09 | support for system call aliasing. stat/fstat/readlink/access etc... gets | Niels Provos | |
grouped into fsread, unlink/rmdir/mkdir goes to fswrite. open switches back between fsread and fswrite depending on oflags parameter. | |||
2002-07-09 | Potential off-by-five, ok provos@ | Daniel Hartmeier | |
2002-07-07 | missing error checks on strdup. from cloder@acm.org | Niels Provos | |
2002-06-28 | KNF | Theo de Raadt | |
2002-06-21 | rewrite all system call arguments in the permit case. use realpath | Niels Provos | |
when we still have the root and we of the monitored process. this eliminates almost all race coniditions. | |||
2002-06-19 | more careful buffer handling; pointed out by deraadt@ | Niels Provos | |
2002-06-10 | support attaching to a running process; some code by fries@ | Niels Provos | |
2002-06-04 | complain about missing device only once. from deraadt@ | Niels Provos | |