linux kernel bug???

[xiaofe...@sina.com]

This is the first time for me to report something, so maybe the format is not right. Please  correct me if I'm wrong. Thank you ~

When using Syzkaller to fuzz Linux kernel 5.11.0, it reported one null-ptr-def bug in file.c. The information is as follows:

The kernel's commit:   c03c21ba6

The syzkaller commit: 52e3731913ab2677be27c29ed8142b04e8f28521

The report:

Syzkaller hit 'KASAN: null-ptr-deref Read in __fget_light' bug.

==================================================================
audit: type=1400 audit(1614571961.594:8): avc:  denied  { execmem } for  pid=294 comm="syz-executor894" scontext=system_u:system_r:kernel_t:s0 tcontext=system_u:system_r:kernel_t:s0 tclass=process permissive=1
BUG: KASAN: null-ptr-deref in instrument_atomic_read include/linux/instrumented.h:71 [inline]
BUG: KASAN: null-ptr-deref in atomic_read include/asm-generic/atomic-instrumented.h:27 [inline]
BUG: KASAN: null-ptr-deref in __fget_light+0x4f/0x250 fs/file.c:930
Read of size 4 at addr 0000000000000000 by task io_wqe_worker-0/296

CPU: 1 PID: 296 Comm: io_wqe_worker-0 Not tainted 5.11.0+ #5
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014
Call Trace:
 __dump_stack lib/dump_stack.c:79 [inline]
 dump_stack+0xaf/0xf2 lib/dump_stack.c:120
 __kasan_report mm/kasan/report.c:400 [inline]
 kasan_report.cold.9+0x10c/0x10e mm/kasan/report.c:413
 check_memory_region_inline mm/kasan/generic.c:179 [inline]
 check_memory_region+0x198/0x200 mm/kasan/generic.c:185
 instrument_atomic_read include/linux/instrumented.h:71 [inline]
 atomic_read include/asm-generic/atomic-instrumented.h:27 [inline]
 __fget_light+0x4f/0x250 fs/file.c:930
 fdget_raw include/linux/file.h:70 [inline]
 path_init+0x940/0x1290 fs/namei.c:2354
 path_lookupat.isra.65+0x2c/0x4e0 fs/namei.c:2410
 filename_lookup.part.80+0x17a/0x370 fs/namei.c:2453
 filename_lookup fs/namei.c:2446 [inline]
 user_path_at_empty+0x4b/0x80 fs/namei.c:2733
 user_path_at include/linux/namei.h:60 [inline]
 vfs_statx+0x108/0x310 fs/stat.c:195
 do_statx+0xa9/0x120 fs/stat.c:590
 io_statx fs/io_uring.c:4525 [inline]
 io_issue_sqe+0x1c31/0x4a50 fs/io_uring.c:6343
 io_wq_submit_work+0x2f7/0x8e0 fs/io_uring.c:6418
 io_worker_handle_work+0xcdd/0x1900 fs/io-wq.c:561
 io_wqe_worker+0xa15/0xeb0 fs/io-wq.c:603
 kthread+0x32a/0x3f0 kernel/kthread.c:292
 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:294
==================================================================


Syzkaller reproducer:
# {Threaded:false Collide:false Repeat:false RepeatTimes:0 Procs:1 Slowdown:1 Sandbox: Fault:false FaultCall:-1 FaultNth:0 Leak:false NetInjection:false NetDevices:false NetReset:false Cgroups:false BinfmtMisc:false CloseFDs:false KCSAN:false DevlinkPCI:false USB:false VhciInjection:false Wifi:false Sysctl:false UseTmpDir:false HandleSegv:false Repro:false Trace:false}
r0 = syz_io_uring_setup(0x1, &(0x7f0000000080)={0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, &(0x7f00000a0000)=nil, &(0x7f00000b0000)=nil, &(0x7f0000000100)=<r1=>0x0, &(0x7f0000000140)=<r2=>0x0)
syz_io_uring_submit(r1, r2, &(0x7f00000008c0)=@IORING_OP_STATX={0x15, 0x0, 0x0, 0xffffffffffffffff, 0x0, &(0x7f00000003c0)='./file1\x00', 0x20, 0x1000}, 0x3)
io_uring_enter(r0, 0x1, 0x1, 0x1, 0x0, 0x0)


C reproducer:
// autogenerated by syzkaller (https://github.com/google/syzkaller)

#define _GNU_SOURCE

#include <endian.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/mman.h>
#include <sys/syscall.h>
#include <sys/types.h>
#include <unistd.h>

#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

struct io_sqring_offsets {
uint32_t head;
uint32_t tail;
uint32_t ring_mask;
uint32_t ring_entries;
uint32_t flags;
uint32_t dropped;
uint32_t array;
uint32_t resv1;
uint64_t resv2;
};

struct io_cqring_offsets {
uint32_t head;
uint32_t tail;
uint32_t ring_mask;
uint32_t ring_entries;
uint32_t overflow;
uint32_t cqes;
uint64_t resv[2];
};

struct io_uring_params {
uint32_t sq_entries;
uint32_t cq_entries;
uint32_t flags;
uint32_t sq_thread_cpu;
uint32_t sq_thread_idle;
uint32_t features;
uint32_t resv[4];
struct io_sqring_offsets sq_off;
struct io_cqring_offsets cq_off;
};

#define IORING_OFF_SQ_RING 0
#define IORING_OFF_SQES 0x10000000ULL

#define sys_io_uring_setup 425
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(sys_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;
}

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;
}

#ifndef __NR_io_uring_enter
#define __NR_io_uring_enter 426
#endif

uint64_t r[3] = {0xffffffffffffffff, 0x0, 0x0};

int main(void)
{
syscall(__NR_mmap, 0x1ffff000ul, 0x1000ul, 0ul, 0x32ul, -1, 0ul);
syscall(__NR_mmap, 0x20000000ul, 0x1000000ul, 7ul, 0x32ul, -1, 0ul);
syscall(__NR_mmap, 0x21000000ul, 0x1000ul, 0ul, 0x32ul, -1, 0ul);
intptr_t res = 0;
*(uint32_t*)0x20000084 = 0;
*(uint32_t*)0x20000088 = 0;
*(uint32_t*)0x2000008c = 0;
*(uint32_t*)0x20000090 = 0;
*(uint32_t*)0x20000098 = 0;
*(uint32_t*)0x2000009c = 0;
*(uint32_t*)0x200000a0 = 0;
*(uint32_t*)0x200000a4 = 0;
res = -1;
res = syz_io_uring_setup(1, 0x20000080, 0x200a0000, 0x200b0000, 0x20000100, 0x20000140);
if (res != -1) {
r[0] = res;
r[1] = *(uint64_t*)0x20000100;
r[2] = *(uint64_t*)0x20000140;
}
*(uint8_t*)0x200008c0 = 0x15;
*(uint8_t*)0x200008c1 = 0;
*(uint16_t*)0x200008c2 = 0;
*(uint32_t*)0x200008c4 = -1;
*(uint64_t*)0x200008c8 = 0;
*(uint64_t*)0x200008d0 = 0x200003c0;
memcpy((void*)0x200003c0, "./file1\000", 8);
*(uint32_t*)0x200008d8 = 0x20;
*(uint32_t*)0x200008dc = 0x1000;
*(uint64_t*)0x200008e0 = 0;
*(uint16_t*)0x200008e8 = 0;
*(uint16_t*)0x200008ea = 0;
*(uint8_t*)0x200008ec = 0;
*(uint8_t*)0x200008ed = 0;
*(uint8_t*)0x200008ee = 0;
*(uint8_t*)0x200008ef = 0;
*(uint8_t*)0x200008f0 = 0;
*(uint8_t*)0x200008f1 = 0;
*(uint8_t*)0x200008f2 = 0;
*(uint8_t*)0x200008f3 = 0;
*(uint8_t*)0x200008f4 = 0;
*(uint8_t*)0x200008f5 = 0;
*(uint8_t*)0x200008f6 = 0;
*(uint8_t*)0x200008f7 = 0;
*(uint8_t*)0x200008f8 = 0;
*(uint8_t*)0x200008f9 = 0;
*(uint8_t*)0x200008fa = 0;
*(uint8_t*)0x200008fb = 0;
*(uint8_t*)0x200008fc = 0;
*(uint8_t*)0x200008fd = 0;
*(uint8_t*)0x200008fe = 0;
*(uint8_t*)0x200008ff = 0;
syz_io_uring_submit(r[1], r[2], 0x200008c0, 3);
syscall(__NR_io_uring_enter, r[0], 1, 1, 1ul, 0ul, 0ul);
return 0;
}

[xiaofe...@sina.com]

[xiaofe...@sina.com]

[Lukas Bulwahn]

Hi,

On Mon, Mar 1, 2021 at 7:58 AM <xiaofe...@sina.com> wrote:
>
> This is the first time for me to report something, so maybe the format is not right. Please correct me if I'm wrong. Thank you ~
> When using Syzkaller to fuzz Linux kernel 5.11.0, it reported one null-ptr-def bug in file.c. The information is as follows:
>
> The kernel's commit: c03c21ba6
> The syzkaller commit: 52e3731913ab2677be27c29ed8142b04e8f28521
>

This looks like a valid kernel issue found with syzkaller.

Given that it is reproducible, I suggest you continue with the following steps:

1. Check if it reproduces on v5.12-rc1. If not, there is already a fix
in v5.12-rc1. Try to identify the specific fix by bisection.

2. Check since when this issue occurs, e.g., does it appear on v5.10, v5.9 etc.

3. Bisect it to a specific commit.

4. Report back the results of your bisections.

Lukas

> --
> You received this message because you are subscribed to the Google Groups "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/syzkaller/20210301065837.0AB115D005DA%40webmail.sinamail.sina.com.cn.

[Dmitry Vyukov]

On Mon, Mar 1, 2021 at 8:10 AM Lukas Bulwahn <lukas....@gmail.com> wrote:
>
> Hi,
>
> On Mon, Mar 1, 2021 at 7:58 AM <xiaofe...@sina.com> wrote:
> >
> > This is the first time for me to report something, so maybe the format is not right. Please correct me if I'm wrong. Thank you ~
> > When using Syzkaller to fuzz Linux kernel 5.11.0, it reported one null-ptr-def bug in file.c. The information is as follows:
> >
> > The kernel's commit: c03c21ba6
> > The syzkaller commit: 52e3731913ab2677be27c29ed8142b04e8f28521
> >
>
> This looks like a valid kernel issue found with syzkaller.
>
> Given that it is reproducible, I suggest you continue with the following steps:
>
> 1. Check if it reproduces on v5.12-rc1. If not, there is already a fix
> in v5.12-rc1. Try to identify the specific fix by bisection.
>
> 2. Check since when this issue occurs, e.g., does it appear on v5.10, v5.9 etc.
>
> 3. Bisect it to a specific commit.
>
> 4. Report back the results of your bisections.

Hi,

For syzkaller bugs it's also a good idea to check if it was already
reported by syzbot or not.
This one seems to be already reported:
https://syzkaller.appspot.com/bug?id=e9579dc4fc313abec952e5811bf48a26fb8d74b4
https://lore.kernel.org/lkml/000000000000ff...@google.com/

> To view this discussion on the web visit https://groups.google.com/d/msgid/syzkaller/CAKXUXMz1CqLAz8ZT4202%2Bh9pNzJU%3DYTN7PExAbaAgpOfqzWXgg%40mail.gmail.com.