summaryrefslogtreecommitdiffstats
path: root/sandbox
Commit message (Collapse)AuthorAgeFilesLines
* Fix compiler error in sandbox util in case exception is enabled.victorw@chromium.org2010-06-222-0/+12
| | | | | | | | | | | | | | | | | Add operator delete that matches the placement form of the operator new. This is required by compiler to generate code to call operator delete in case the object's constructor throws an exception. See http://msdn.microsoft.com/en-us/library/cxdxz3x6.aspx R=rvargas BUG=none TEST=sandbox code compiles if exception is enabled. Review URL: http://codereview.chromium.org/2870017 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@50486 0039d316-1c4b-4281-b951-d872f2087c98
* FAILS tests should not turn the bot red.vandebo@chromium.org2010-06-181-2/+2
| | | | | | | | | | | | | | | | | | | Revert 50258 - TTF: Reenable some disabled tests. DevToolsSanityTest.TestConsoleLog seems to work ProxyResolverV8Test.FAILS_ReturnUnicode DepTest.FAILS_TestDepDisable VectorCanvasTest.FAILS_Matrix TEST=none BUG=disabled tests Review URL: http://codereview.chromium.org/2819011 TBR=vandebo@chromium.org Review URL: http://codereview.chromium.org/2862016 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@50261 0039d316-1c4b-4281-b951-d872f2087c98
* TTF: Reenable some disabled tests.vandebo@chromium.org2010-06-181-2/+2
| | | | | | | | | | | | | | DevToolsSanityTest.TestConsoleLog seems to work ProxyResolverV8Test.FAILS_ReturnUnicode DepTest.FAILS_TestDepDisable VectorCanvasTest.FAILS_Matrix TEST=none BUG=disabled tests Review URL: http://codereview.chromium.org/2819011 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@50258 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 49982 - patchthakis@chromium.org2010-06-161-2/+2
| | | | | | | TBR=thakis@chromium.org Review URL: http://codereview.chromium.org/2825006 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@50002 0039d316-1c4b-4281-b951-d872f2087c98
* patchthakis@chromium.org2010-06-161-2/+2
| | | | git-svn-id: svn://svn.chromium.org/chrome/trunk/src@49982 0039d316-1c4b-4281-b951-d872f2087c98
* Rename Chrome threads to use a "Cr" prefix.evan@chromium.org2010-06-091-1/+1
| | | | | | | | | | | Before we carefully used the app (Chrome/Chromium) name, but: 1) these are for internal debugging use only, so why compute this name 2) on Linux the thread ids are limited to 16 characters, and "Chromium" is too long Review URL: http://codereview.chromium.org/2741003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@49296 0039d316-1c4b-4281-b951-d872f2087c98
* Use DISALLOW_COPY_AND_ASSIGN instead of DISALLOW_EVIL_CONSTRUCTORS in ↵tfarina@chromium.org2010-06-0523-51/+51
| | | | | | | | | | | sandbox directory. BUG=None TEST=trybots Review URL: http://codereview.chromium.org/2643003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@49010 0039d316-1c4b-4281-b951-d872f2087c98
* Implement most of the ridealong fixes/cleanups I suggested during review for ↵pkasting@chromium.org2010-06-012-9/+9
| | | | | | | | | | enabling warn-on-signed-versus-unsigned-equality-comparisions on Windows. BUG=none TEST=none Review URL: http://codereview.chromium.org/2395001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48666 0039d316-1c4b-4281-b951-d872f2087c98
* Enable warning 4389 as an error on windows builds. This will make mbelshe@chromium.org2010-05-275-39/+39
| | | | | | | | | | | | windows builds more similar to linux/mac, which already treat signed/ unsigned equality comparisons as warnings (and hence errors). BUG=44471 TEST=none Review URL: http://codereview.chromium.org/2222002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48395 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Another dependency the bbot missed!!!"rafaelw@chromium.org2010-05-265-39/+39
| | | | | | | | | | | | | | | | | | This reverts commit 161f7fd3bdd425167af9fe26fdc5373a2ff44c98. Revert "Missed a file as part of checkin for r48186" This reverts commit cff86beba5938209393a6c3bccced62a7f3ff36b. Revert "Enable warning 4389 as an error on windows builds. This will make" This reverts commit c78936bcfc65b98edf288191d927a495b0364621. TBR=mbelshe Review URL: http://codereview.chromium.org/2253001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48238 0039d316-1c4b-4281-b951-d872f2087c98
* Enable warning 4389 as an error on windows builds. This will makembelshe@chromium.org2010-05-255-39/+39
| | | | | | | | | | | | windows builds more similar to linux/mac, which already treat signed/ unsigned equality comparisons as warnings (and hence errors). BUG=44471 TEST=none Review URL: http://codereview.chromium.org/2081007 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48186 0039d316-1c4b-4281-b951-d872f2087c98
* Seccomp sandbox: Add Gyp rules for building the test suitemseaborn@chromium.org2010-05-241-0/+30
| | | | | | | | | | | | | | This should work both standalone and inside the Chromium build. I have not included an action for running the tests, since having such an action does not seem to be common in the Chromium build. BUG=none TEST=seccomp_tests Review URL: http://codereview.chromium.org/2165001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48043 0039d316-1c4b-4281-b951-d872f2087c98
* Add NEWNS and NEWNET to the SUID sandbox.agl@chromium.org2010-05-241-15/+30
| | | | | | | | | | | | | | | | | | | This patch attempts to fork off the sandboxed process with the additional NEWNS and NEWNET flags. If these flags aren't supported at runtime then the code will degrade to the current behaviour. NEWNS starts children in a new mount namespace so that they cannot affect the parent's mounts. (This is a little bit useless every little helps.) NEWNET starts children in a new network space, initially with no network devices and this stops sandboxed processes from talking to the network. Additionally, children exist in their own namespaces for UNIX domain sockets and the abstract namespace. http://codereview.chromium.org/2108020/show git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48040 0039d316-1c4b-4281-b951-d872f2087c98
* Remove a possible race in the SUID sandbox (minor)agl@chromium.org2010-05-201-7/+18
| | | | | | | | | | | | | | | | | The SUID sandbox can be used to set the oom_adj value for non-dumpable processes owned by the same user. When doing so, we previously first checked the directory owner and then opened the oom_adj file. In between the check and the open, the process could have died and another process could have taken that PID value. We would then adjust the OOM value of the wrong process. Given how PIDs are allocated, this is very hard to exploit and, even then, a minor security issue at best, but we can avoid the issue entirely with openat. http://codereview.chromium.org/2118007 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47801 0039d316-1c4b-4281-b951-d872f2087c98
* seccomp sandbox: Split source file list into a separate .gyp filemseaborn@chromium.org2010-05-202-52/+67
| | | | | | | | | | | | | | | | | The intention behind this is to make it easier to sync the .gyp file into the non-Chromium copy of the seccomp sandbox so that it can be used to build a standalone version of the sandbox. Also, it arguably makes the .gyp files more manageable. Removes a dependency on "base", which the seccomp sandbox does not use. BUG=none TEST=none Review URL: http://codereview.chromium.org/1939002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47792 0039d316-1c4b-4281-b951-d872f2087c98
* Add support for calling {rt_,}sigaction(), and for invoking signal handlersmarkus@chromium.org2010-05-1810-66/+625
| | | | | | | | | | | | | | from within the sandbox. Added tests for the new functionality and merged the tests for sigreturn() that had previously been committed to the standalone version of the sandbox (on Google Code) TEST=run "make test" BUG=37728 Review URL: http://codereview.chromium.org/2074003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47561 0039d316-1c4b-4281-b951-d872f2087c98
* seccomp sandbox: Add tests for signal handlingmseaborn@chromium.org2010-05-181-3/+145
| | | | | | | | | | | | | | | | | | | Test that signal handlers can be run OK. This tests the support for sigreturn() (that involves patching the VDSO) that was added in r76 of the non-Chromium version of the sandbox. Test that signal masks can be set and read. This tests the sigprocmask() support that was added in r70. Add a mechanism for checking that a test exits with an expected non-zero exit status, such as SIGSEGV. BUG=none TEST=test_syscalls Review URL: http://codereview.chromium.org/2087013 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47541 0039d316-1c4b-4281-b951-d872f2087c98
* GYP changes for FreeBSD and OpenBSDpvalchev@google.com2010-05-141-0/+9
| | | | | | Review URL: http://codereview.chromium.org/1480002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47327 0039d316-1c4b-4281-b951-d872f2087c98
* Prevent FPU state corruption by directly creating the signal frame on themarkus@chromium.org2010-05-114-69/+56
| | | | | | | | | | | | | | stack of the newly created thread, instead of creating it on the caller's stack and copying it over. This eliminates the need to do complicated touch-ups of the signal stack's data structure, which turned out to be incorrect for the FPU state. Thanks to Mark Seaborn for pointing out this simplification of the code. TEST=Chrome no longer crashes in tcmalloc BUG=none Review URL: http://codereview.chromium.org/2051005 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@46928 0039d316-1c4b-4281-b951-d872f2087c98
* Add a couple more testsmarkus@chromium.org2010-04-281-7/+126
| | | | | | | | | | | | | | | | | | These tests were useful for debugging reference_trusted_thread.cc. Test an easily-forwarded system call, dup(). Also test clone() directly, in addition to testing it indirectly via pthread_create(). Check for leaked FDs. Change the test runner to run all tests, even if one fails, rather than stopping at the first failed test. Review URL: http://codereview.chromium.org/1750014 BUG=none TEST=none Review URL: http://codereview.chromium.org/1756015 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@45806 0039d316-1c4b-4281-b951-d872f2087c98
* Correctly align memory allocations in the case where we need to allocate memorymarkus@chromium.org2010-04-281-1/+4
| | | | | | | | | | | as far away from the stack as possible, but still as close to the VDSO as we can. BUG=none TEST=run the tests in a tight loop and notice that they no longer randomly fail Review URL: http://codereview.chromium.org/1807002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@45775 0039d316-1c4b-4281-b951-d872f2087c98
* Added support for sigreturn() and rt_sigreturn(). On x86-32, this ismarkus@chromium.org2010-04-285-127/+239
| | | | | | | | | | | | | | | | | | | | | | | | | complicated by the fact that in Seccomp mode, we can only ever call sigreturn(). But in order to eventually support sigaction(), we want to be able to also call rt_sigreturn(). We solve this problem by rewriting the signal stack frame from an RT signal frame to a legacy frame. Fortunately, this part of the signal frame is stable between kernel versions. The unstable part (i.e. extended registers such as FP, MMX, SSE, ...) is always identical in both in both types of signal frames. None of these complications exist on x86-64 and it is relatively straight-forward to enable support for the system call. The only difficulty lies in the fact that its calling conventions are somewhat different from "normal" system calls. So, we have to handle rt_sigreturn() from within the syscallWrapper() and the segv() handler and cannot write it in C code. TEST=ad hoc testing until we have support for sigaction(). Then we can add a unittest BUG=37728 Review URL: http://codereview.chromium.org/1739011 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@45774 0039d316-1c4b-4281-b951-d872f2087c98
* Merged Mark Seaborn's changes:markus@chromium.org2010-04-2218-76/+427
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add some automated tests for seccomp-sandbox This covers some of the syscalls that are proxied by the trusted process. This adds a basic test runner framework. Fix error return code for open() and other syscalls on x86-64 On error, the open() syscall was returning errno & 0xffffffff in %rax on x86-64 instead of -errno. This stops glibc from regarding this as an error, and so its syscall wrapper returns -errno instead of -1 and does not set errno. I have fixed up the other syscalls to use long (64-bit) instead of int (32-bit) as well. Not all of them were affected by the problem: it depends on gcc's code generation. Sometimes casting to int truncates a value, sometimes it doesn't. It seems better to be consistent though. Adds a test for open() and some other syscalls. TODO: Need to figure out how to run the tests automatically BUG=none TEST=none Review URL: http://codereview.chromium.org/1729003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@45379 0039d316-1c4b-4281-b951-d872f2087c98
* Make the use of signals inside of the sandbox safe.markus@chromium.org2010-04-2012-409/+965
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We previously assumed that no signals would ever be enabled in the sandbox and thus there was no way to trick the trusted thread into executing potentially untrusted code. In an attempt to lift this restriction, this changelist modifies the trusted thread so that - it has an invalid stack pointer at all times. Any attempt to handle a signal would result in the kernel trying to push a signal stack, which would immediately result in a SEGV and then terminate the application. - all signals are blocked while outside of trusted code. If a signal is triggered, it either gets handled on one of the sandboxed threads (for asynchronous signals), or it results in the application getting terminated by the kernel (for synchronous signals). This changelist is difficult not only because eliminating all uses of the stack pointer requires some very careful assembly coding, but more importantly because we have to restore signals after we enter seccomp mode. As sigprocmask() is a restricted system call, the only way to restore the signal mask is by calling sigreturn() with a suitably tweaked signal stack frame. While the first couple of bytes of the signal stack frame are well-defined and unlikely to change, the entire signal stack frame is not documented as part of the stable ABI. The exact format depends on the number of modified CPU registers (e.g. SSE, MMX, floating point, ...) The only way for us to get a valid signal stack frame is to trigger a signal, and to create a (possibly adjusted) copy of the signal frame. We obviously have to do this _before_ we block all signals upon entering trusted code. The two places where this needs to happen is upon start of the sandbox when launching the initial trusted thread, and upon any call to clone(). BUG=37728 TEST=Run chrome and verify that /proc/$PID/status shows the correct signal mask for trusted threads. The latter can be identified with strace. Review URL: http://codereview.chromium.org/1594040 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@45055 0039d316-1c4b-4281-b951-d872f2087c98
* SELinux: add basic policy.agl@chromium.org2010-04-193-0/+53
| | | | | | | | | | | | | | | | This patch removes the chromium_zygote_t type and adds a chromium_renderer_t type. Also, a basic policy for chromium_renderer_t is included. I decided not to try to have a different policy for the zygote since it just makes things more complex for little reason. BUG=none TEST=none http://codereview.chromium.org/1104002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@44908 0039d316-1c4b-4281-b951-d872f2087c98
* Fix SELinux warnings when running on Fedora.markus@chromium.org2010-03-301-3/+3
| | | | | | | | | | | | (c.f. http://people.redhat.com/drepper/selinux-mem.html) Fix compilation warnings on Fedora. BUG=none TEST=when running Chrome on Fedora, verify that we don't get AVC warnings Review URL: http://codereview.chromium.org/1535004 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@43107 0039d316-1c4b-4281-b951-d872f2087c98
* Update/add copyright headers.markus@chromium.org2010-03-293-30/+13
| | | | | | | | BUG=32501 TEST=none Review URL: http://codereview.chromium.org/1574001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@42992 0039d316-1c4b-4281-b951-d872f2087c98
* linux: turn on -Wextraevan@chromium.org2010-03-267-22/+22
| | | | | | | | | | | | | | This is a followup to an earlier change (r38266) which did most of the warning-related cleanup. This one just flips the flag, and fixes some new warnings that crept in during the window while the flag was off. Second try, now with some libpng fixes. BUG=34160 Review URL: http://codereview.chromium.org/1320011 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@42700 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "linux: turn on -Wextra"evan@chromium.org2010-03-257-22/+22
| | | | | | | | Compiled locally and on trybots but failed on builder?! This reverts commit r42688. git-svn-id: svn://svn.chromium.org/chrome/trunk/src@42689 0039d316-1c4b-4281-b951-d872f2087c98
* linux: turn on -Wextraevan@chromium.org2010-03-257-22/+22
| | | | | | | | | | | | This is a followup to an earlier change (r38266) which did most of the warning-related cleanup. This one just flips the flag, and fixes some new warnings that crept in during the window while the flag was off. BUG=34160 Review URL: http://codereview.chromium.org/597023 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@42688 0039d316-1c4b-4281-b951-d872f2087c98
* Fix a few more places where we need to use our own allocator.markus@chromium.org2010-03-252-5/+19
| | | | | | | | | | Make tcmalloc compatible with the seccomp sandbox by avoiding making direct system calls from within tcmalloc. BUG=38973 TEST=none Review URL: http://codereview.chromium.org/1294001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@42667 0039d316-1c4b-4281-b951-d872f2087c98
* Fix the POC and some of unit tests to be build andnsylvain@chromium.org2010-03-236-34/+30
| | | | | | | | | | run correctly in 64 bit. BUG=27218 Review URL: http://codereview.chromium.org/1168002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@42392 0039d316-1c4b-4281-b951-d872f2087c98
* Add #include to fix compile errors on "Linux Perf (webkit.org)" bot.yutak@chromium.org2010-03-181-0/+1
| | | | | | | | | | TBR=markus BUG=none TEST=none Review URL: http://codereview.chromium.org/1107001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41931 0039d316-1c4b-4281-b951-d872f2087c98
* Add a first version of a tool for analyzing performance data output by themarkus@chromium.org2010-03-182-0/+197
| | | | | | | | | | sandbox. BUG=none TEST=none Review URL: http://codereview.chromium.org/1076001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41917 0039d316-1c4b-4281-b951-d872f2087c98
* Compute and pring the time that it takes to execute system calls. This datamarkus@chromium.org2010-03-1820-84/+327
| | | | | | | | | | | | is going to be skewed slightly, as calling gettimeofday() by itself also takes a little bit of time. But it should be good enough to allow us to see where we have performance bottlenecks. TEST=none BUG=none Review URL: http://codereview.chromium.org/997009 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41905 0039d316-1c4b-4281-b951-d872f2087c98
* Marked UnloadDllTest.UnloadAviCapDllWithPatching as flaky.glider@chromium.org2010-03-171-1/+1
| | | | | | | | TBR=cpu Review URL: http://codereview.chromium.org/1002009 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41832 0039d316-1c4b-4281-b951-d872f2087c98
* Third change to port sandbox to 64 bitscpu@chromium.org2010-03-165-35/+28
| | | | | | | | | | | | - Enable registry, file and event policy - Now sbox_integration_tests.exe runs and all tests pass BUG=27218 TEST=included Review URL: http://codereview.chromium.org/992003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41673 0039d316-1c4b-4281-b951-d872f2087c98
* Re-land second sandbox round of changescpu@chromium.org2010-03-1320-172/+351
| | | | | | | | | | | | | | | | | | | | | - broke integration tests See http://codereview.chromium.org/757001 BUG=27218 TEST=included No change in the original files. The only changes are in the files: sandbox/src/registry_dispatcher.cc sandbox/src/registry_dispatcher.h sandbox/src/filesystem_dispatcher.cc Because now even in 32 bits the handles map to void pointers Review URL: http://codereview.chromium.org/873006 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41523 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 41481 - Secound round of sbox changes for 64 bit portericu@google.com2010-03-1217-340/+155
| | | | | | | | | | | | | | | | | | Handling pointer sized items Beefing up unit tests Beefing up integration tests Enabling Process, Thread and Token IPCs Making validation tests compile again BUG=27218 TEST= unit tests included Review URL: http://codereview.chromium.org/757001 TBR=cpu@chromium.org Review URL: http://codereview.chromium.org/871008 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41491 0039d316-1c4b-4281-b951-d872f2087c98
* Secound round of sbox changes for 64 bit portcpu@chromium.org2010-03-1217-155/+340
| | | | | | | | | | | | | | | - Handling pointer sized items - Beefing up unit tests - Beefing up integration tests - Enabling Process, Thread and Token IPCs - Making validation tests compile again BUG=27218 TEST= unit tests included Review URL: http://codereview.chromium.org/757001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41481 0039d316-1c4b-4281-b951-d872f2087c98
* Fix warning that is now an error.craig.schlenter@chromium.org2010-03-101-2/+1
| | | | | | | | C:\b\slave\chromium-rel-xp\build\src\sandbox\src\sandbox.cc(16) : warning C4010: single-line comment contains line-continuation character Review URL: http://codereview.chromium.org/788003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41164 0039d316-1c4b-4281-b951-d872f2087c98
* Make sandbox code compile as "chromium_code".craig.schlenter@chromium.org2010-03-102-2/+6
| | | | | | | | | | | | | | | | This sets up useful flags like -Wall -Werror etc. Also squash a compiler warning: sandbox/linux/suid/process_util_linux.c: In function ‘AdjustOOMScore’: sandbox/linux/suid/process_util_linux.c:25: error: format ‘%lu’ expects type ‘long unsigned int’, but argument 4 has type ‘pid_t’ BUG=none TEST=try-servers Review URL: http://codereview.chromium.org/733001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@41161 0039d316-1c4b-4281-b951-d872f2087c98
* Added missing copyright header.markus@chromium.org2010-03-0822-0/+88
| | | | | | | | BUG=32501 TEST=none Review URL: http://codereview.chromium.org/672011 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@40946 0039d316-1c4b-4281-b951-d872f2087c98
* - Add a custom allocator for STL objects. This fixes sandbox failures thatmarkus@chromium.org2010-03-0813-218/+643
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | were observed on some machines (in particular in 32bit mode). - Some more changes to avoid calling into glibc when we can make a direct system call, instead. These particular call sites were unlikely to cause any problems. But it makes the code easier to audit if we avoid all unnecessary calls into glibc. - In 64bit mode, gettimeofday() is handled by vsyscalls and tends to be cheap. In 32bit mode, it is just a regular system call. Some users rely on being able to call gettimeofday() at a very high rate (up to thousands of consecutive calls). Recognize this system call pattern and optimize for it. - Add debugging option that allows us to warn about expensive system calls. In many cases, these warnings can then be used to optimize the sandboxed application. - Fix compilation on newer versions of gcc. - Changed the x86-32 version of the code that we use when intercepting system calls. Previously, we would use CALL to jump to the set of instructions that we had relocated. But we made the mistake of allowing relocation of instructions that reference %esp. This doesn't work, as CALL modifies the stack. We now avoid using CALL and instead jump directly. On x86-32 that requires the use of a PUSH/RET combination as there is no 32bit wide JMP instruction. The x86-64 version of the code was already written in a way that would avoid this particular problem. (I would like to thank Craig Schlenter for his exceptional detective work in tracking down the root cause of this bug!) - For debugging purposes, injected a really small library (less than 4kB) and discovered that some of our memory map manipulations implicitly relied on mappings to be at least two pages long. Fixed the code that made this incorrect assumption. - For really small libraries, the runtime linker can choose a different more compact layout. Our computation of the ASR offset did not know how to deal with that. Fixed by explicitly looking for a ".text" segment instead of looking for a PT_DYNAMIC section. - Closed a file descriptor that we kept open longer than needed. - Removed some unused code. - Added copyright headers TEST=tested on i386 and x86-64 BUG=36133 Review URL: http://codereview.chromium.org/661438 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@40900 0039d316-1c4b-4281-b951-d872f2087c98
* Sandbox: Add code for all the interceptors.rvargas@google.com2010-03-055-14/+317
| | | | | | | | | | | | Still waiting for IPC to be able to test anything. BUG=27218 TEST=none Review URL: http://codereview.chromium.org/669128 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@40700 0039d316-1c4b-4281-b951-d872f2087c98
* Porting sbox IPC to 64 bitcpu@chromium.org2010-03-039-49/+65
| | | | | | | | | | | | | | | | | -IPC unit tests enabled -Integration IPCTest test enabled -All unit test pass (62 tests) Other tests require interceptor fu. BUG=27218 TEST= unit tests included Review URL: http://codereview.chromium.org/661299 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@40485 0039d316-1c4b-4281-b951-d872f2087c98
* Be more restrictive when finding file names for libraries that need patching.markus@chromium.org2010-02-241-2/+17
| | | | | | | | | | | | | | | | | | | | | This avoids false positives if the directory name matches one of the well-known library names (e.g. ld). False positives not only result in a performance hit at startup, because we are now trying to instrument libraries that don't actually contain any system calls; but even worse than this, we could try to instrument system calls in the sandboxing code itself. And those system calls are deliberately coded so that they will not get rewritten. Fortunately, none of this is a security problem. If we accidentally rewrite system calls that weren't supposed to be rewritten, we will just crash on startup. TEST=the sandbox now works on the buildbots BUG=36133 Review URL: http://codereview.chromium.org/652188 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39839 0039d316-1c4b-4281-b951-d872f2087c98
* Explicitly ask for unsigned values when comparing addresses. Not only is thismarkus@chromium.org2010-02-241-2/+2
| | | | | | | | | | | code hard to understand (and possibly broken) otherwise, some versions of GCC complain about the comparison without the cast. TEST=none BUG=none Review URL: http://codereview.chromium.org/657034 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39838 0039d316-1c4b-4281-b951-d872f2087c98
* Treat calls to lstat() and lstat64() the same as calls to stat(). In practise,markus@chromium.org2010-02-243-17/+72
| | | | | | | | | | | | | | | this means the calls will still be denied. But we now return a correct return code. But more importantly, this change brings the source code in line with the code of the stand-alone opensource sandbox. Wherever possible, we try to keep both code bases identical. TEST=none BUG=none Review URL: http://codereview.chromium.org/657040 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39837 0039d316-1c4b-4281-b951-d872f2087c98
* If /tmp is not a POSIX file system, try to use /dev/shm for creating ourmarkus@chromium.org2010-02-231-3/+58
| | | | | | | | | | | temporary directory. BUG=30926 TEST=tested with tmpfs, ext3 and NFS Review URL: http://codereview.chromium.org/650177 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@39679 0039d316-1c4b-4281-b951-d872f2087c98
generated by cgit v1.1 at 2025-01-18 05:07:48 +0000