Re: KASAN: slab-out-of-bounds Read in ld_usb_read (3)

[Johan Hovold]

On Wed, Oct 16, 2019 at 06:42:10PM -0700, syzbot wrote:
> Hello,
> 
> syzbot found the following crash on:
> 
> HEAD commit:    22be26f7 usb-fuzzer: main usb gadget fuzzer driver
> git tree:       https://github.com/google/kasan.git usb-fuzzer
> console output: https://syzkaller.appspot.com/x/log.txt?x=14f6dc5f600000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=387eccb7ac68ec5
> dashboard link: https://syzkaller.appspot.com/bug?extid=6fe95b826644f7f12b0b
> compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=102c3227600000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=12a503a0e00000
> 
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: syzbot+6fe95b826644f7f12b0b@xxxxxxxxxxxxxxxxxxxxxxxxx
> 
> ldusb 1-1:0.28: Read buffer overflow, -131383859965943 bytes dropped

This warning should be using %zu and thus read 18446612689849585673.

Apparently, *actual_buffer holds a kernel address ffff8881cfb42000
instead of a transfer length. Possibly a stale value since the buffer is
not cleared on allocation.

> ==================================================================
> BUG: KASAN: slab-out-of-bounds in _copy_to_user+0x124/0x150  
> lib/usercopy.c:28
> Read of size 102391 at addr ffff8881cfb40008 by task syz-executor372/1737

And due to missing sanity checks the driver proceeds to access data
beyond the ring-buffer entry.

> CPU: 0 PID: 1737 Comm: syz-executor372 Not tainted 5.4.0-rc3+ #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS  
> Google 01/01/2011
> Call Trace:
>   __dump_stack lib/dump_stack.c:77 [inline]
>   dump_stack+0xca/0x13e lib/dump_stack.c:113
>   print_address_description.constprop.0+0x36/0x50 mm/kasan/report.c:374
>   __kasan_report.cold+0x1a/0x33 mm/kasan/report.c:506
>   kasan_report+0xe/0x20 mm/kasan/common.c:634
>   check_memory_region_inline mm/kasan/generic.c:185 [inline]
>   check_memory_region+0x128/0x190 mm/kasan/generic.c:192
>   _copy_to_user+0x124/0x150 lib/usercopy.c:28
>   copy_to_user include/linux/uaccess.h:152 [inline]
>   ld_usb_read+0x329/0x760 drivers/usb/misc/ldusb.c:492
>   __vfs_read+0x76/0x100 fs/read_write.c:425
>   vfs_read+0x1ea/0x430 fs/read_write.c:461
>   ksys_read+0x1e8/0x250 fs/read_write.c:587
>   do_syscall_64+0xb7/0x580 arch/x86/entry/common.c:290
>   entry_SYSCALL_64_after_hwframe+0x49/0xbe
> RIP: 0033:0x4421d9
> Code: e8 7c e7 ff ff 48 83 c4 18 c3 0f 1f 80 00 00 00 00 48 89 f8 48 89 f7  
> 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff  
> ff 0f 83 bb 07 fc ff c3 66 2e 0f 1f 84 00 00 00 00
> RSP: 002b:00007ffc08eb0888 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
> RAX: ffffffffffffffda RBX: 00007ffc08eb0ae0 RCX: 00000000004421d9
> RDX: 0000000000018ff7 RSI: 0000000020000a80 RDI: 0000000000000004
> RBP: 0000000000000000 R08: 000000000000000f R09: 0000000000402eb0
> R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
> R13: 0000000000402eb0 R14: 0000000000000000 R15: 0000000000000000
> 
> The buggy address belongs to the page:
> page:ffffea00073ed000 refcount:1 mapcount:0 mapping:0000000000000000  
> index:0x0 compound_mapcount: 0
> flags: 0x200000000010000(head)
> raw: 0200000000010000 dead000000000100 dead000000000122 0000000000000000
> raw: 0000000000000000 0000000000000000 00000001ffffffff 0000000000000000
> page dumped because: kasan: bad access detected
> 
> Memory state around the buggy address:
>   ffff8881cfb55500: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>   ffff8881cfb55580: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> > ffff8881cfb55600: fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe
>                     ^
>   ffff8881cfb55680: fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe
>   ffff8881cfb55700: fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe fe
> ==================================================================

The drivers ring-buffer implementation is broken so this could
potentially be a memory-ordering issue (e.g. the entry is read before it
has been updated).

Let's try the below.

Johan


#syz test: https://github.com/google/kasan.git 22be26f7