| Message ID | 20200827052659.24922-4-cmr@codefail.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Use per-CPU temporary mappings for patching | expand |
Hi "Christopher,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on powerpc/next]
[also build test ERROR on char-misc/char-misc-testing tip/x86/core v5.9-rc2 next-20200827]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Christopher-M-Riedl/Use-per-CPU-temporary-mappings-for-patching/20200827-161532
base: https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git next
config: x86_64-allmodconfig (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0
reproduce (this is a W=1 build):
# save the attached .config to linux build tree
make W=1 ARCH=x86_64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
All errors (new ones prefixed by >>):
drivers/misc/lkdtm/perms.c: In function 'lkdtm_HIJACK_PATCH':
>> drivers/misc/lkdtm/perms.c:318:38: error: implicit declaration of function 'read_cpu_patching_addr' [-Werror=implicit-function-declaration]
318 | addr = offset_in_page(patch_site) | read_cpu_patching_addr(patching_cpu);
| ^~~~~~~~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
# https://github.com/0day-ci/linux/commit/36a98d779ee4620e6e091cbe3b438b52faa108ad
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Christopher-M-Riedl/Use-per-CPU-temporary-mappings-for-patching/20200827-161532
git checkout 36a98d779ee4620e6e091cbe3b438b52faa108ad
vim +/read_cpu_patching_addr +318 drivers/misc/lkdtm/perms.c
289
290 void lkdtm_HIJACK_PATCH(void)
291 {
292 #ifdef CONFIG_PPC
293 struct ppc_inst original_insn = ppc_inst_read(READ_ONCE(patch_site));
294 #endif
295 #ifdef CONFIG_X86_64
296 int original_insn = READ_ONCE(*patch_site);
297 #endif
298 struct task_struct *patching_kthrd;
299 int patching_cpu, hijacker_cpu, attempts;
300 unsigned long addr;
301 bool hijacked;
302 const int bad_data = 0xbad00bad;
303
304 if (num_online_cpus() < 2) {
305 pr_warn("need at least two cpus\n");
306 return;
307 }
308
309 hijacker_cpu = smp_processor_id();
310 patching_cpu = cpumask_any_but(cpu_online_mask, hijacker_cpu);
311
312 patching_kthrd = kthread_create_on_node(&lkdtm_patching_cpu, NULL,
313 cpu_to_node(patching_cpu),
314 "lkdtm_patching_cpu");
315 kthread_bind(patching_kthrd, patching_cpu);
316 wake_up_process(patching_kthrd);
317
> 318 addr = offset_in_page(patch_site) | read_cpu_patching_addr(patching_cpu);
319
320 pr_info("starting hijacker_cpu=%d\n", hijacker_cpu);
321 for (attempts = 0; attempts < 100000; ++attempts) {
322 /* Use __put_user to catch faults without an Oops */
323 hijacked = !__put_user(bad_data, (int *)addr);
324
325 if (hijacked) {
326 if (kthread_stop(patching_kthrd))
327 pr_err("error trying to stop patching thread\n");
328 break;
329 }
330 }
331 pr_info("hijack attempts: %d\n", attempts);
332
333 if (hijacked) {
334 if (lkdtm_verify_patch(bad_data))
335 pr_err("overwrote kernel text\n");
336 /*
337 * There are window conditions where the hijacker cpu manages to
338 * write to the patch site but the site gets overwritten again by
339 * the patching cpu. We still consider that a "successful" hijack
340 * since the hijacker cpu did not fault on the write.
341 */
342 pr_err("FAIL: wrote to another cpu's patching area\n");
343 } else {
344 kthread_stop(patching_kthrd);
345 }
346
347 /* Restore the original insn for any future lkdtm tests */
348 #ifdef CONFIG_PPC
349 patch_instruction(patch_site, original_insn);
350 #endif
351 #ifdef CONFIG_X86_64
352 lkdtm_do_patch(original_insn);
353 #endif
354 }
355
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
Hi "Christopher,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on powerpc/next]
[also build test ERROR on char-misc/char-misc-testing tip/x86/core v5.9-rc2 next-20200827]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Christopher-M-Riedl/Use-per-CPU-temporary-mappings-for-patching/20200827-161532
base: https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git next
config: parisc-allyesconfig (attached as .config)
compiler: hppa-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=parisc
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
All errors (new ones prefixed by >>):
>> hppa-linux-ld: drivers/misc/lkdtm/core.o:(.rodata+0x1b4): undefined reference to `lkdtm_HIJACK_PATCH'
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
diff --git a/drivers/misc/lkdtm/core.c b/drivers/misc/lkdtm/core.c index a5e344df9166..482e72f6a1e1 100644 --- a/drivers/misc/lkdtm/core.c +++ b/drivers/misc/lkdtm/core.c @@ -145,6 +145,7 @@ static const struct crashtype crashtypes[] = { CRASHTYPE(WRITE_RO), CRASHTYPE(WRITE_RO_AFTER_INIT), CRASHTYPE(WRITE_KERN), + CRASHTYPE(HIJACK_PATCH), CRASHTYPE(REFCOUNT_INC_OVERFLOW), CRASHTYPE(REFCOUNT_ADD_OVERFLOW), CRASHTYPE(REFCOUNT_INC_NOT_ZERO_OVERFLOW), diff --git a/drivers/misc/lkdtm/lkdtm.h b/drivers/misc/lkdtm/lkdtm.h index 8878538b2c13..8bd98e8f0443 100644 --- a/drivers/misc/lkdtm/lkdtm.h +++ b/drivers/misc/lkdtm/lkdtm.h @@ -60,6 +60,7 @@ void lkdtm_EXEC_USERSPACE(void); void lkdtm_EXEC_NULL(void); void lkdtm_ACCESS_USERSPACE(void); void lkdtm_ACCESS_NULL(void); +void lkdtm_HIJACK_PATCH(void); /* lkdtm_refcount.c */ void lkdtm_REFCOUNT_INC_OVERFLOW(void); diff --git a/drivers/misc/lkdtm/perms.c b/drivers/misc/lkdtm/perms.c index 2dede2ef658f..0ed32aba5216 100644 --- a/drivers/misc/lkdtm/perms.c +++ b/drivers/misc/lkdtm/perms.c @@ -9,6 +9,7 @@ #include <linux/vmalloc.h> #include <linux/mman.h> #include <linux/uaccess.h> +#include <linux/kthread.h> #include <asm/cacheflush.h> /* Whether or not to fill the target memory area with do_nothing(). */ @@ -222,6 +223,151 @@ void lkdtm_ACCESS_NULL(void) pr_err("FAIL: survived bad write\n"); } +#if defined(CONFIG_PPC) || defined(CONFIG_X86_64) +#if defined(CONFIG_STRICT_KERNEL_RWX) && defined(CONFIG_SMP) +/* + * This is just a dummy location to patch-over. + */ +static void patching_target(void) +{ + return; +} + +#ifdef CONFIG_PPC +#include <asm/code-patching.h> +struct ppc_inst * const patch_site = (struct ppc_inst *)&patching_target; +#endif + +#ifdef CONFIG_X86_64 +#include <asm/text-patching.h> +int * const patch_site = (int *)&patching_target; +#endif + +static inline int lkdtm_do_patch(int data) +{ +#ifdef CONFIG_PPC + return patch_instruction(patch_site, ppc_inst(data)); +#endif +#ifdef CONFIG_X86_64 + text_poke(patch_site, &data, sizeof(int)); + return 0; +#endif +} + +static inline bool lkdtm_verify_patch(int data) +{ +#ifdef CONFIG_PPC + return ppc_inst_equal(ppc_inst_read(READ_ONCE(patch_site)), + ppc_inst(data)); +#endif +#ifdef CONFIG_X86_64 + return READ_ONCE(*patch_site) == data; +#endif +} + +static int lkdtm_patching_cpu(void *data) +{ + int err = 0; + int val = 0xdeadbeef; + + pr_info("starting patching_cpu=%d\n", smp_processor_id()); + do { + err = lkdtm_do_patch(val); + } while (lkdtm_verify_patch(val) && !err && !kthread_should_stop()); + + if (err) + pr_warn("patch_instruction returned error: %d\n", err); + + set_current_state(TASK_INTERRUPTIBLE); + while (!kthread_should_stop()) { + schedule(); + set_current_state(TASK_INTERRUPTIBLE); + } + + return err; +} + +void lkdtm_HIJACK_PATCH(void) +{ +#ifdef CONFIG_PPC + struct ppc_inst original_insn = ppc_inst_read(READ_ONCE(patch_site)); +#endif +#ifdef CONFIG_X86_64 + int original_insn = READ_ONCE(*patch_site); +#endif + struct task_struct *patching_kthrd; + int patching_cpu, hijacker_cpu, attempts; + unsigned long addr; + bool hijacked; + const int bad_data = 0xbad00bad; + + if (num_online_cpus() < 2) { + pr_warn("need at least two cpus\n"); + return; + } + + hijacker_cpu = smp_processor_id(); + patching_cpu = cpumask_any_but(cpu_online_mask, hijacker_cpu); + + patching_kthrd = kthread_create_on_node(&lkdtm_patching_cpu, NULL, + cpu_to_node(patching_cpu), + "lkdtm_patching_cpu"); + kthread_bind(patching_kthrd, patching_cpu); + wake_up_process(patching_kthrd); + + addr = offset_in_page(patch_site) | read_cpu_patching_addr(patching_cpu); + + pr_info("starting hijacker_cpu=%d\n", hijacker_cpu); + for (attempts = 0; attempts < 100000; ++attempts) { + /* Use __put_user to catch faults without an Oops */ + hijacked = !__put_user(bad_data, (int *)addr); + + if (hijacked) { + if (kthread_stop(patching_kthrd)) + pr_err("error trying to stop patching thread\n"); + break; + } + } + pr_info("hijack attempts: %d\n", attempts); + + if (hijacked) { + if (lkdtm_verify_patch(bad_data)) + pr_err("overwrote kernel text\n"); + /* + * There are window conditions where the hijacker cpu manages to + * write to the patch site but the site gets overwritten again by + * the patching cpu. We still consider that a "successful" hijack + * since the hijacker cpu did not fault on the write. + */ + pr_err("FAIL: wrote to another cpu's patching area\n"); + } else { + kthread_stop(patching_kthrd); + } + + /* Restore the original insn for any future lkdtm tests */ +#ifdef CONFIG_PPC + patch_instruction(patch_site, original_insn); +#endif +#ifdef CONFIG_X86_64 + lkdtm_do_patch(original_insn); +#endif +} + +#else + +void lkdtm_HIJACK_PATCH(void) +{ + if (!IS_ENABLED(CONFIG_PPC) && !IS_ENABLED(CONFIG_X86_64)) + pr_err("XFAIL: this test only runs on x86_64 or powerpc\n"); + if (!IS_ENABLED(CONFIG_STRICT_KERNEL_RWX)) + pr_err("XFAIL: this test requires CONFIG_STRICT_KERNEL_RWX\n"); + if (!IS_ENABLED(CONFIG_SMP)) + pr_err("XFAIL: this test requires CONFIG_SMP\n"); +} + +#endif /* CONFIG_STRICT_KERNEL_RWX && CONFIG_SMP */ +#endif /* CONFIG_PPC || CONFIG_X86_64 */ + void __init lkdtm_perms_init(void) { /* Make sure we can write to __ro_after_init values during __init */
When live patching with STRICT_KERNEL_RWX, the CPU doing the patching must temporarily remap the page(s) containing the patch site with +W permissions. While this temporary mapping is in use another CPU could write to the same mapping and maliciously alter kernel text. Implement a LKDTM test to attempt to exploit such an opening from another (ie. not the patching) CPU. The test is implemented on x86_64 and powerpc only. The LKDTM "hijack" test works as follows: 1. A CPU executes an infinite loop to patch an instruction. This is the "patching" CPU. 2. Another CPU attempts to write to the address of the temporary mapping used by the "patching" CPU. This other CPU is the "hijacker" CPU. The hijack either fails with a segfault or succeeds, in which case some kernel text is now overwritten. How to run the test: mount -t debugfs none /sys/kernel/debug (echo HIJACK_PATCH > /sys/kernel/debug/provoke-crash/DIRECT) Signed-off-by: Christopher M. Riedl <cmr@codefail.de> --- drivers/misc/lkdtm/core.c | 1 + drivers/misc/lkdtm/lkdtm.h | 1 + drivers/misc/lkdtm/perms.c | 146 +++++++++++++++++++++++++++++++++++++ 3 files changed, 148 insertions(+)