^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * This code fills the used part of the kernel stack with a poison value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * before returning to userspace. It's part of the STACKLEAK feature
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * ported from grsecurity/PaX.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * Author: Alexander Popov <alex.popov@linux.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) * STACKLEAK reduces the information which kernel stack leak bugs can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * reveal and blocks some uninitialized stack variable attacks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include <linux/stackleak.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/kprobes.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #ifdef CONFIG_STACKLEAK_RUNTIME_DISABLE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <linux/jump_label.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <linux/sysctl.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) static DEFINE_STATIC_KEY_FALSE(stack_erasing_bypass);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) int stack_erasing_sysctl(struct ctl_table *table, int write,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) void *buffer, size_t *lenp, loff_t *ppos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) int state = !static_branch_unlikely(&stack_erasing_bypass);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) int prev_state = state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) table->data = &state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) table->maxlen = sizeof(int);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) state = !!state;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) if (ret || !write || state == prev_state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) if (state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) static_branch_disable(&stack_erasing_bypass);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) static_branch_enable(&stack_erasing_bypass);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) pr_warn("stackleak: kernel stack erasing is %s\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) state ? "enabled" : "disabled");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) #define skip_erasing() static_branch_unlikely(&stack_erasing_bypass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) #define skip_erasing() false
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) #endif /* CONFIG_STACKLEAK_RUNTIME_DISABLE */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) asmlinkage void noinstr stackleak_erase(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) /* It would be nice not to have 'kstack_ptr' and 'boundary' on stack */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) unsigned long kstack_ptr = current->lowest_stack;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) unsigned long boundary = (unsigned long)end_of_stack(current);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) unsigned int poison_count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) const unsigned int depth = STACKLEAK_SEARCH_DEPTH / sizeof(unsigned long);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) if (skip_erasing())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) /* Check that 'lowest_stack' value is sane */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) if (unlikely(kstack_ptr - boundary >= THREAD_SIZE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) kstack_ptr = boundary;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) /* Search for the poison value in the kernel stack */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) while (kstack_ptr > boundary && poison_count <= depth) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) if (*(unsigned long *)kstack_ptr == STACKLEAK_POISON)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) poison_count++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) poison_count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) kstack_ptr -= sizeof(unsigned long);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) * One 'long int' at the bottom of the thread stack is reserved and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) * should not be poisoned (see CONFIG_SCHED_STACK_END_CHECK=y).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) if (kstack_ptr == boundary)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) kstack_ptr += sizeof(unsigned long);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) #ifdef CONFIG_STACKLEAK_METRICS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) current->prev_lowest_stack = kstack_ptr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) * Now write the poison value to the kernel stack. Start from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) * 'kstack_ptr' and move up till the new 'boundary'. We assume that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) * the stack pointer doesn't change when we write poison.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) if (on_thread_stack())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) boundary = current_stack_pointer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) boundary = current_top_of_stack();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) while (kstack_ptr < boundary) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) *(unsigned long *)kstack_ptr = STACKLEAK_POISON;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) kstack_ptr += sizeof(unsigned long);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) /* Reset the 'lowest_stack' value for the next syscall */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) current->lowest_stack = current_top_of_stack() - THREAD_SIZE/64;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) void __used __no_caller_saved_registers noinstr stackleak_track_stack(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) unsigned long sp = current_stack_pointer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) * Having CONFIG_STACKLEAK_TRACK_MIN_SIZE larger than
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) * STACKLEAK_SEARCH_DEPTH makes the poison search in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) * stackleak_erase() unreliable. Let's prevent that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) BUILD_BUG_ON(CONFIG_STACKLEAK_TRACK_MIN_SIZE > STACKLEAK_SEARCH_DEPTH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) /* 'lowest_stack' should be aligned on the register width boundary */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) sp = ALIGN(sp, sizeof(unsigned long));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) if (sp < current->lowest_stack &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) sp >= (unsigned long)task_stack_page(current) +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) sizeof(unsigned long)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) current->lowest_stack = sp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) EXPORT_SYMBOL(stackleak_track_stack);