gems-kernel/source/THIRDPARTY/xnu/bsd/kern/kern_debug.c
2024-06-03 11:29:39 -05:00

569 lines
16 KiB
C

/*
* Copyright (c) 2020 Apple Inc. All rights reserved.
*
* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
*
* This file contains Original Code and/or Modifications of Original Code
* as defined in and that are subject to the Apple Public Source License
* Version 2.0 (the 'License'). You may not use this file except in
* compliance with the License. The rights granted to you under the License
* may not be used to create, or enable the creation or redistribution of,
* unlawful or unlicensed copies of an Apple operating system, or to
* circumvent, violate, or enable the circumvention or violation of, any
* terms of an Apple operating system software license agreement.
*
* Please obtain a copy of the License at
* http://www.opensource.apple.com/apsl/ and read it before using this file.
*
* The Original Code and all software distributed under the License are
* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
* INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
* Please see the License for the specific language governing rights and
* limitations under the License.
*
* @APPLE_OSREFERENCE_LICENSE_HEADER_END@
*/
#include <sys/sysctl.h>
#include <kern/cpu_data.h>
#if __arm64__
#include <arm/machine_routines.h>
#endif /* __arm64__ */
#if CONFIG_DEBUG_SYSCALL_REJECTION
#include <mach/mach_time.h>
#include <kern/bits.h>
#include <kern/clock.h>
#include <kern/exc_guard.h>
#include <kern/exception.h>
#include <kern/kalloc.h>
#include <kern/simple_lock.h>
#include <kern/startup.h>
#include <kern/syscall_sw.h>
#include <kern/task.h>
#include <pexpert/pexpert.h>
#include <sys/syscall.h>
#include <sys/sysent.h>
#include <sys/systm.h>
#include <sys/types.h>
#include <sys/user.h>
#include <sys/variant_internal.h>
#include <sys/kern_debug.h>
#define SYSCALL_REJECTION_MODE_IGNORE 0
#define SYSCALL_REJECTION_MODE_GUARD 1
#define SYSCALL_REJECTION_MODE_CRASH 2
TUNABLE_WRITEABLE(int, debug_syscall_rejection_mode, "syscall_rejection_mode",
#if DEVELOPMENT || DEBUG
SYSCALL_REJECTION_MODE_GUARD
#else
SYSCALL_REJECTION_MODE_IGNORE
#endif
);
static int
sysctl_debug_syscall_rejection_mode(struct sysctl_oid __unused *oidp, void * __unused arg1, int __unused arg2,
struct sysctl_req *req)
{
int error, changed;
int value = *(int *) arg1;
if (!os_variant_has_internal_diagnostics("com.apple.xnu")) {
return ENOTSUP;
}
error = sysctl_io_number(req, value, sizeof(value), &value, &changed);
if (!error && changed) {
debug_syscall_rejection_mode = value;
}
return error;
}
void
reset_debug_syscall_rejection_mode(void)
{
if (!os_variant_has_internal_diagnostics("com.apple.xnu")) {
debug_syscall_rejection_mode = 0;
}
}
SYSCTL_PROC(_kern, OID_AUTO, debug_syscall_rejection_mode, CTLFLAG_RW | CTLFLAG_LOCKED,
&debug_syscall_rejection_mode, 0, sysctl_debug_syscall_rejection_mode, "I", "0: ignore, 1: non-fatal, 2: crash");
static size_t const predefined_masks = 2; // 0: null mask (all 0), 1: all mask (all 1)
/*
* The number of masks is derived from the mask selector width:
*
* A selector is just made of an index into syscall_rejection_masks,
* with the exception of the highest bit, which indicates whether the
* mask is to be added as an "allow" mask or a "deny" mask.
* Additionally, predefined masks don't actually have storage and are
* handled specially, so syscall_rejection_masks starts with the first
* non-predefined mask (and is sized appropriately).
*/
static size_t const syscall_rejection_mask_count = SYSCALL_REJECTION_SELECTOR_MASK_COUNT - predefined_masks;
static syscall_rejection_mask_t syscall_rejection_masks[syscall_rejection_mask_count];
#define SR_MASK_SIZE (BITMAP_SIZE(mach_trap_count + nsysent))
static LCK_GRP_DECLARE(syscall_rejection_lck_grp, "syscall rejection lock");
static LCK_MTX_DECLARE(syscall_rejection_mtx, &syscall_rejection_lck_grp);
bool
debug_syscall_rejection_handle(int syscall_mach_trap_number)
{
uthread_t ut = current_uthread();
uint64_t const flags = ut->syscall_rejection_flags;
bool fatal = (bool)(flags & SYSCALL_REJECTION_FLAGS_FORCE_FATAL);
switch (debug_syscall_rejection_mode) {
case SYSCALL_REJECTION_MODE_IGNORE:
if (!fatal) {
/* ignore */
break;
}
OS_FALLTHROUGH;
case SYSCALL_REJECTION_MODE_CRASH:
fatal = true;
OS_FALLTHROUGH;
case SYSCALL_REJECTION_MODE_GUARD: {
if (flags & SYSCALL_REJECTION_FLAGS_ONCE) {
int const number = syscall_mach_trap_number < 0 ? -syscall_mach_trap_number : (mach_trap_count + syscall_mach_trap_number);
// don't trip on this system call again
bitmap_set(ut->syscall_rejection_mask, number);
bitmap_set(ut->syscall_rejection_once_mask, number);
}
mach_exception_code_t code = 0;
EXC_GUARD_ENCODE_TYPE(code, GUARD_TYPE_REJECTED_SC);
EXC_GUARD_ENCODE_FLAVOR(code, 0);
EXC_GUARD_ENCODE_TARGET(code, syscall_mach_trap_number < 0);
mach_exception_subcode_t subcode =
syscall_mach_trap_number < 0 ? -syscall_mach_trap_number : syscall_mach_trap_number;
if (!fatal) {
task_violated_guard(code, subcode, NULL, TRUE);
} else {
thread_guard_violation(current_thread(), code, subcode, fatal);
}
break;
};
default:
/* ignore */
;
}
return fatal;
}
extern int exit_with_guard_exception(void *p, mach_exception_data_type_t code,
mach_exception_data_type_t subcode);
void
rejected_syscall_guard_ast(
thread_t t,
mach_exception_data_type_t code,
mach_exception_data_type_t subcode)
{
const bool fatal = true;
/*
* Check if anyone has registered for Synchronous EXC_GUARD, if yes then,
* deliver it synchronously and then kill the process, else kill the process
* and deliver the exception via EXC_CORPSE_NOTIFY. Always kill the process if we are not in dev mode.
*/
if (task_exception_notify(EXC_GUARD, code, subcode, fatal) == KERN_SUCCESS) {
psignal_uthread(t, SIGSYS);
} else {
exit_with_guard_exception(current_proc(), code, subcode);
}
}
static void
_syscall_rejection_apply_mask(syscall_rejection_mask_t dest, const syscall_rejection_mask_t src, bool apply_as_allow)
{
assert(dest != NULL);
assert(src != NULL);
if (apply_as_allow) {
bitmap_or(dest, dest, src, mach_trap_count + nsysent);
} else {
bitmap_and_not(dest, dest, src, mach_trap_count + nsysent);
}
}
/*
* The masks to apply are passed to the kernel as packed selectors,
* which are just however many of the selector data type fit into one
* (or more) fields of the natural word size (i.e. a register). This
* avoids copying from user space.
*
* More specifically, at the time of this writing, a selector is 7
* bits wide, and there are two uint64_t arguments
* (args->packed_selectors<n>), so up to 18 selectors can be
* specified, which are then stuffed into the 128 bits of the
* arguments. If less than 18 masks are requested to be applied, the
* remaining selectors will just be left as 0, which naturally
* resolves as the "empty" or "NULL" mask that changes nothing.
*
* The libsyscall wrapper provides a more convenient interface where
* an array (up to 18 elements long) and its length are passed in,
* which the wrapper then packs into packed_selectors of the actual
* system call.
*/
int
sys_debug_syscall_reject_config(struct proc *p __unused, struct debug_syscall_reject_config_args *args, int *retval)
{
int error = 0;
*retval = 0;
uthread_t ut = current_uthread();
bitmap_t mask[SR_MASK_SIZE / sizeof(bitmap_t)];
// syscall rejection masks are always reset to "deny all"
memset(mask, 0, SR_MASK_SIZE);
lck_mtx_lock(&syscall_rejection_mtx);
for (int i = 0;
i + SYSCALL_REJECTION_SELECTOR_BITS < (sizeof(args->packed_selectors1) + sizeof(args->packed_selectors2)) * 8;
i += SYSCALL_REJECTION_SELECTOR_BITS) {
#define s_left_shift(x, n) ((n) < 0 ? ((x) >> -(n)) : ((x) << (n)))
syscall_rejection_selector_t const selector = (syscall_rejection_selector_t)
(((i < 64 ? (args->packed_selectors1 >> i) : 0) |
(i > 64 - SYSCALL_REJECTION_SELECTOR_BITS ? s_left_shift(args->packed_selectors2, 64 - i) : 0)) & SYSCALL_REJECTION_SELECTOR_MASK);
bool const is_allow_mask = selector & SYSCALL_REJECTION_IS_ALLOW_MASK;
int const mask_index = selector & SYSCALL_REJECTION_INDEX_MASK;
if (mask_index == SYSCALL_REJECTION_NULL) {
// mask 0 is always empty (nothing to apply)
continue;
}
if (mask_index == SYSCALL_REJECTION_ALL) {
// mask 1 is always full (overrides everything)
memset(mask, is_allow_mask ? 0xff : 0x00, SR_MASK_SIZE);
continue;
}
syscall_rejection_mask_t mask_to_apply = syscall_rejection_masks[mask_index - predefined_masks];
if (mask_to_apply == NULL) {
error = ENOENT;
goto out_locked;
}
_syscall_rejection_apply_mask(mask, mask_to_apply, is_allow_mask);
}
/* Not RT-safe, but only necessary once. */
if (ut->syscall_rejection_mask == NULL) {
ut->syscall_rejection_mask = kalloc_data(SR_MASK_SIZE, Z_WAITOK);
if (ut->syscall_rejection_mask == NULL) {
error = ENOMEM;
goto out_locked;
}
}
memcpy(ut->syscall_rejection_mask, mask, SR_MASK_SIZE);
if ((args->flags & SYSCALL_REJECTION_FLAGS_ONCE)) {
if (ut->syscall_rejection_once_mask == NULL) {
ut->syscall_rejection_once_mask = kalloc_data(SR_MASK_SIZE, Z_WAITOK);
if (ut->syscall_rejection_once_mask == NULL) {
kfree_data(ut->syscall_rejection_mask, SR_MASK_SIZE);
ut->syscall_rejection_mask = NULL;
error = ENOMEM;
goto out_locked;
}
memset(ut->syscall_rejection_once_mask, 0, SR_MASK_SIZE);
} else {
// prevent the already hit syscalls from hitting again.
bitmap_or(ut->syscall_rejection_mask, ut->syscall_rejection_mask, ut->syscall_rejection_once_mask, mach_trap_count + nsysent);
}
}
out_locked:
lck_mtx_unlock(&syscall_rejection_mtx);
if (error == 0) {
ut->syscall_rejection_flags = args->flags;
}
if (error == ENOENT && debug_syscall_rejection_mode == SYSCALL_REJECTION_MODE_IGNORE) {
/* Existing code may rely on the system call failing
* gracefully if syscall rejection is currently off. */
error = 0;
}
return error;
}
/*
* debug_syscall_reject
*
* Compatibility interface to the old form of the system call.
*/
int
debug_syscall_reject(struct proc *p, struct debug_syscall_reject_args *args, int *retval)
{
struct debug_syscall_reject_config_args new_args;
bzero(&new_args, sizeof(new_args));
new_args.packed_selectors1 = args->packed_selectors;
// packed_selectors2 left empty
new_args.flags = SYSCALL_REJECTION_FLAGS_DEFAULT;
return sys_debug_syscall_reject_config(p, &new_args, retval);
}
static bool
_syscall_rejection_add(syscall_rejection_mask_t dst, char const *name)
{
/*
* Yes, this function is O(n+m), making the whole act of setting a
* mask O(l*(n+m)), but defining masks is done rarely enough (and
* i, n and m small enough) for this to not matter.
*/
for (int i = 0; i < mach_trap_count; i++) {
if (strcmp(mach_syscall_name_table[i], name) == 0) {
bitmap_set(dst, i);
return true;
}
}
extern char const *syscallnames[];
for (int i = 0; i < nsysent; i++) {
if (strcmp(syscallnames[i], name) == 0) {
bitmap_set(dst, i + mach_trap_count);
return true;
}
}
printf("%s: trying to add non-existing syscall/mach trap '%s'\n", __func__, name);
return false;
}
/* Pretty much arbitrary, we just don't want userspace to pass
* unreasonably large buffers to parse. */
static size_t const max_input_size = 16 * PAGE_MAX_SIZE;
static int
_sysctl_debug_syscall_rejection_masks(struct sysctl_oid __unused *oidp, void * __unused arg1, int __unused arg2,
struct sysctl_req *req)
{
size_t const max_name_len = 128;
char name[max_name_len];
if (req->newptr == 0) {
return 0;
}
if (req->newlen > max_input_size) {
return E2BIG;
}
size_t const len = req->newlen;
char *buf = kalloc_data(len + 1, Z_WAITOK);
if (buf == NULL) {
return ENOMEM;
}
/*
* sysctl_io_string always copies out the given buffer as the
* "old" value if requested. We could construct a text
* representation of existing masks, but this is not particularly
* interesting, so we just return the dummy string "<masks>".
*/
strlcpy(buf, "<masks>", len + 1);
int changed = 0;
int error = sysctl_io_string(req, buf, len + 1, 0, &changed);
if (error != 0 || !changed) {
goto out;
}
char const *p = buf;
int id = 0;
int l = 0;
int n = sscanf(p, "%i: %n", &id, &l);
if (n != 1 || id < predefined_masks || id > syscall_rejection_mask_count + predefined_masks) {
printf("%s: invalid mask id %i (or conversion failed)\n", __FUNCTION__, id);
error = EINVAL;
goto out;
}
p += l;
syscall_rejection_mask_t new_mask = kalloc_data(SR_MASK_SIZE,
Z_WAITOK | Z_ZERO);
if (new_mask == NULL) {
printf("%s: allocating new mask for id %i failed\n", __FUNCTION__, id);
error = ENOMEM;
goto out;
}
error = 0;
while (p < buf + len && *p != 0) {
name[0] = 0;
n = sscanf(p, "%127s %n", name, &l);
if (n != 1 || name[0] == 0) {
error = EINVAL;
kfree_data(new_mask, SR_MASK_SIZE);
goto out;
}
if (!_syscall_rejection_add(new_mask, name)) {
error = ENOENT;
kfree_data(new_mask, SR_MASK_SIZE);
goto out;
}
p += l;
}
syscall_rejection_mask_t to_free = NULL;
lck_mtx_lock(&syscall_rejection_mtx);
syscall_rejection_mask_t *target_mask = &syscall_rejection_masks[id - predefined_masks];
to_free = *target_mask;
*target_mask = new_mask;
lck_mtx_unlock(&syscall_rejection_mtx);
kfree_data(to_free, SR_MASK_SIZE);
out:
kfree_data(buf, len + 1);
return error;
}
SYSCTL_PROC(_kern, OID_AUTO, syscall_rejection_masks, CTLTYPE_STRING | CTLFLAG_WR | CTLFLAG_MASKED | CTLFLAG_LOCKED,
0, 0, _sysctl_debug_syscall_rejection_masks, "A", "system call rejection masks");
#else /* CONFIG_DEBUG_SYSCALL_REJECTION */
#include <sys/kern_debug.h>
int
sys_debug_syscall_reject_config(struct proc * __unused p, struct debug_syscall_reject_config_args * __unused args, int __unused *ret)
{
/* not supported. */
return ENOTSUP;
}
int
debug_syscall_reject(struct proc * __unused p, struct debug_syscall_reject_args * __unused args, int * __unused retval)
{
/* not supported. */
return ENOTSUP;
}
void
reset_debug_syscall_rejection_mode(void)
{
/* not supported. */
}
#endif /* CONFIG_DEBUG_SYSCALL_REJECTION */
#if __arm64__ && (DEBUG || DEVELOPMENT)
static void
_spinfor(uint64_t nanoseconds)
{
uint64_t mt = 0;
nanoseconds_to_absolutetime(nanoseconds, &mt);
uint64_t start = mach_absolute_time();
while (mach_absolute_time() < start + mt) {
// Spinning.
}
}
static int
_sysctl_debug_disable_interrupts_test(struct sysctl_oid __unused *oidp, void * __unused arg1, int __unused arg2,
struct sysctl_req *req)
{
int error = 0;
if (req->newptr == 0) {
goto out;
}
uint64_t val = 0;
error = sysctl_io_number(req, 0, sizeof(val), &val, NULL);
if (error != 0 || val == 0) {
goto out;
}
boolean_t istate = ml_set_interrupts_enabled(false);
_spinfor(val);
ml_set_interrupts_enabled(istate);
out:
return error;
}
static int
_sysctl_debug_disable_preemption_test(struct sysctl_oid __unused *oidp, void * __unused arg1, int __unused arg2,
struct sysctl_req *req)
{
int error = 0;
if (req->newptr == 0) {
goto out;
}
uint64_t val = 0;
error = sysctl_io_number(req, 0, sizeof(val), &val, NULL);
if (error != 0 || val == 0) {
goto out;
}
disable_preemption();
_spinfor(val);
enable_preemption();
out:
return error;
}
SYSCTL_PROC(_kern, OID_AUTO, debug_disable_interrupts_test, CTLTYPE_QUAD | CTLFLAG_WR | CTLFLAG_MASKED | CTLFLAG_LOCKED,
0, 0, _sysctl_debug_disable_interrupts_test, "Q", "disable interrupts for specified number of nanoseconds, for testing");
SYSCTL_PROC(_kern, OID_AUTO, debug_disable_preemption_test, CTLTYPE_QUAD | CTLFLAG_WR | CTLFLAG_MASKED | CTLFLAG_LOCKED,
0, 0, _sysctl_debug_disable_preemption_test, "Q", "disable preemption for specified number of nanoseconds, for testing");
#endif /* __arm64__ && (DEBUG || DEVELOPMENT) */