#
73f62e60 |
|
03-Mar-2022 |
Kees Cook <keescook@chromium.org> |
lkdtm: Move crashtype definitions into each category It's long been annoying that to add a new LKDTM test one had to update lkdtm.h and core.c to get it "registered". Switch to a per-category list and update the crashtype walking code in core.c to handle it. This also means that all the lkdtm_* tests themselves can be static now. Cc: Arnd Bergmann <arnd@arndb.de> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kees Cook <keescook@chromium.org>
|
#
5e5a6c54 |
|
15-Feb-2022 |
Christophe Leroy <christophe.leroy@csgroup.eu> |
lkdtm: Add a test for function descriptors protection Add WRITE_OPD to check that you can't modify function descriptors. Gives the following result when function descriptors are not protected: lkdtm: Performing direct entry WRITE_OPD lkdtm: attempting bad 16 bytes write at c00000000269b358 lkdtm: FAIL: survived bad write lkdtm: do_nothing was hijacked! Looks like a standard compiler barrier() is not enough to force GCC to use the modified function descriptor. Had to add a fake empty inline assembly to force GCC to reload the function descriptor. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/7eeba50d16a35e9d799820e43304150225f20197.1644928018.git.christophe.leroy@csgroup.eu
|
#
72a86433 |
|
15-Feb-2022 |
Christophe Leroy <christophe.leroy@csgroup.eu> |
lkdtm: Fix execute_[user]_location() execute_location() and execute_user_location() intent to copy do_nothing() text and execute it at a new location. However, at the time being it doesn't copy do_nothing() function but do_nothing() function descriptor which still points to the original text. So at the end it still executes do_nothing() at its original location allthough using a copied function descriptor. So, fix that by really copying do_nothing() text and build a new function descriptor by copying do_nothing() function descriptor and updating the target address with the new location. Also fix the displayed addresses by dereferencing do_nothing() function descriptor. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/4055839683d8d643cd99be121f4767c7c611b970.1644928018.git.christophe.leroy@csgroup.eu
|
#
b6491339 |
|
15-Feb-2022 |
Christophe Leroy <christophe.leroy@csgroup.eu> |
lkdtm: Really write into kernel text in WRITE_KERN WRITE_KERN is supposed to overwrite some kernel text, namely do_overwritten() function. But at the time being it overwrites do_overwritten() function descriptor, not function text. Fix it by dereferencing the function descriptor to obtain function text pointer. Export dereference_function_descriptor() for when LKDTM is built as a module. And make do_overwritten() noinline so that it is really do_overwritten() which is called by lkdtm_WRITE_KERN(). Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/31e58eaffb5bc51c07d8d4891d1982100ade8cfc.1644928018.git.christophe.leroy@csgroup.eu
|
#
69b420ed |
|
15-Feb-2022 |
Christophe Leroy <christophe.leroy@csgroup.eu> |
lkdtm: Force do_nothing() out of line LKDTM tests display that the run do_nothing() at a given address, but in reality do_nothing() is inlined into the caller. Force it out of line so that it really runs text at the displayed address. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by: Kees Cook <keescook@chromium.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/a5dcf4d2088e6aca47ab3b4c6d5c0f7fa064e25a.1644928018.git.christophe.leroy@csgroup.eu
|
#
464e86b4 |
|
25-Jun-2020 |
Kees Cook <keescook@chromium.org> |
lkdtm: Avoid more compiler optimizations for bad writes It seems at least Clang is able to throw away writes it knows are destined for read-only memory, which makes things like the WRITE_RO test fail, as the write gets elided. Instead, force the variable to be volatile, and make similar changes through-out other tests in an effort to avoid needing to repeat fixing these kinds of problems. Also includes pr_err() calls in failure paths so that kernel logs are more clear in the failure case. Reported-by: Prasad Sodagudi <psodagud@codeaurora.org> Suggested-by: Sami Tolvanen <samitolvanen@google.com> Fixes: 9ae113ce5faf ("lkdtm: add tests for additional page permissions") Signed-off-by: Kees Cook <keescook@chromium.org> Link: https://lore.kernel.org/r/20200625203704.317097-2-keescook@chromium.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
59a12205 |
|
14-Dec-2018 |
Christophe Leroy <christophe.leroy@c-s.fr> |
lkdtm: Add tests for NULL pointer dereference Introduce lkdtm tests for NULL pointer dereference: check access or exec at NULL address, since these errors tend to be reported differently from the general fault error text. For example from x86: pr_alert("BUG: unable to handle kernel %s at %px\n", address < PAGE_SIZE ? "NULL pointer dereference" : "paging request", (void *)address); Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr> Signed-off-by: Kees Cook <keescook@chromium.org>
|
#
4c411157 |
|
07-Nov-2018 |
Christophe Leroy <christophe.leroy@c-s.fr> |
lkdtm: Print real addresses Today, when doing a lkdtm test before the readiness of the random generator, (ptrval) is printed instead of the address at which it perform the fault: [ 1597.337030] lkdtm: Performing direct entry EXEC_USERSPACE [ 1597.337142] lkdtm: attempting ok execution at (ptrval) [ 1597.337398] lkdtm: attempting bad execution at (ptrval) [ 1597.337460] kernel tried to execute user page (77858000) -exploit attempt? (uid: 0) [ 1597.344769] Unable to handle kernel paging request for instruction fetch [ 1597.351392] Faulting instruction address: 0x77858000 [ 1597.356312] Oops: Kernel access of bad area, sig: 11 [#1] If the lkdtm test is done later on, it prints an hashed address. In both cases this is pointless. The purpose of the test is to ensure the kernel generates an Oops at the expected address, so real addresses needs to be printed. This patch fixes that. Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr> Signed-off-by: Kees Cook <keescook@chromium.org>
|
#
039a1c42 |
|
06-Mar-2018 |
Kees Cook <keescook@chromium.org> |
lkdtm: Relocate code to subdirectory The LKDTM modules keep expanding, and it's getting weird to have each file get a prefix. Instead, move to a subdirectory for cleaner handling. Signed-off-by: Kees Cook <keescook@chromium.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|