Re: [PATCH bpf] bpf, test_run: Fix use-after-free issue in eth_skb_pkt_type()

[Martin KaFai Lau <martin.lau@linux.dev>]

On 12/2/24 1:38 PM, Martin KaFai Lau wrote:
> On 12/2/24 8:15 AM, Stanislav Fomichev wrote:
>> On 12/02, Shigeru Yoshida wrote:
>>> KMSAN reported a use-after-free issue in eth_skb_pkt_type()[1]. The
>>> cause of the issue was that eth_skb_pkt_type() accessed skb's data
>>> that didn't contain an Ethernet header. This occurs when
>>> bpf_prog_test_run_xdp() passes an invalid value as the user_data
>>> argument to bpf_test_init().
>>>
>>> Fix this by returning an error when user_data is less than ETH_HLEN in
>>> bpf_test_init().
>>>
>>> [1]
>>> BUG: KMSAN: use-after-free in eth_skb_pkt_type include/linux/ 
>>> etherdevice.h:627 [inline]
>>> BUG: KMSAN: use-after-free in eth_type_trans+0x4ee/0x980 net/ethernet/eth.c:165
>>>   eth_skb_pkt_type include/linux/etherdevice.h:627 [inline]
>>>   eth_type_trans+0x4ee/0x980 net/ethernet/eth.c:165
>>>   __xdp_build_skb_from_frame+0x5a8/0xa50 net/core/xdp.c:635
>>>   xdp_recv_frames net/bpf/test_run.c:272 [inline]
>>>   xdp_test_run_batch net/bpf/test_run.c:361 [inline]
>>>   bpf_test_run_xdp_live+0x2954/0x3330 net/bpf/test_run.c:390
>>>   bpf_prog_test_run_xdp+0x148e/0x1b10 net/bpf/test_run.c:1318
>>>   bpf_prog_test_run+0x5b7/0xa30 kernel/bpf/syscall.c:4371
>>>   __sys_bpf+0x6a6/0xe20 kernel/bpf/syscall.c:5777
>>>   __do_sys_bpf kernel/bpf/syscall.c:5866 [inline]
>>>   __se_sys_bpf kernel/bpf/syscall.c:5864 [inline]
>>>   __x64_sys_bpf+0xa4/0xf0 kernel/bpf/syscall.c:5864
>>>   x64_sys_call+0x2ea0/0x3d90 arch/x86/include/generated/asm/syscalls_64.h:322
>>>   do_syscall_x64 arch/x86/entry/common.c:52 [inline]
>>>   do_syscall_64+0xd9/0x1d0 arch/x86/entry/common.c:83
>>>   entry_SYSCALL_64_after_hwframe+0x77/0x7f
>>>
>>> Uninit was created at:
>>>   free_pages_prepare mm/page_alloc.c:1056 [inline]
>>>   free_unref_page+0x156/0x1320 mm/page_alloc.c:2657
>>>   __free_pages+0xa3/0x1b0 mm/page_alloc.c:4838
>>>   bpf_ringbuf_free kernel/bpf/ringbuf.c:226 [inline]
>>>   ringbuf_map_free+0xff/0x1e0 kernel/bpf/ringbuf.c:235
>>>   bpf_map_free kernel/bpf/syscall.c:838 [inline]
>>>   bpf_map_free_deferred+0x17c/0x310 kernel/bpf/syscall.c:862
>>>   process_one_work kernel/workqueue.c:3229 [inline]
>>>   process_scheduled_works+0xa2b/0x1b60 kernel/workqueue.c:3310
>>>   worker_thread+0xedf/0x1550 kernel/workqueue.c:3391
>>>   kthread+0x535/0x6b0 kernel/kthread.c:389
>>>   ret_from_fork+0x6e/0x90 arch/x86/kernel/process.c:147
>>>   ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
>>>
>>> CPU: 1 UID: 0 PID: 17276 Comm: syz.1.16450 Not tainted 6.12.0-05490- 
>>> g9bb88c659673 #8
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-3.fc41 
>>> 04/01/2014
>>>
>>> Fixes: be3d72a2896c ("bpf: move user_size out of bpf_test_init")
>>> Reported-by: syzkaller <syzkaller@googlegroups.com>
>>> Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
>>> ---
>>>   net/bpf/test_run.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
>>> index 501ec4249fed..756250aa890f 100644
>>> --- a/net/bpf/test_run.c
>>> +++ b/net/bpf/test_run.c
>>> @@ -663,7 +663,7 @@ static void *bpf_test_init(const union bpf_attr *kattr, 
>>> u32 user_size,
>>>       if (size < ETH_HLEN || size > PAGE_SIZE - headroom - tailroom)
>>>           return ERR_PTR(-EINVAL);
>>> -    if (user_size > size)
>>> +    if (user_size < ETH_HLEN || user_size > size)
>>>           return ERR_PTR(-EMSGSIZE);
>>>       size = SKB_DATA_ALIGN(size);
>>> -- 
>>> 2.47.0
>>>
>>
>> I wonder whether 'size < ETH_HLEN' above is needed after your patch.
>> Feels like 'user_size < ETH_HLEN' supersedes it.
> 
> May be fixing it by replacing the existing "size" check with "user_size" check? 
> Seems more intuitive that checking is needed on the "user_"size instead of the 
> "size". The "if (user_size > size)" check looks useless also. Something like this?
> 
> -    if (size < ETH_HLEN || size > PAGE_SIZE - headroom - tailroom)
> +    if (user_size < ETH_HLEN || user_size > PAGE_SIZE - headroom - tailroom)
>          return ERR_PTR(-EINVAL);
> 
> -    if (user_size > size)
> -        return ERR_PTR(-EMSGSIZE);
> -

just noticed the xdp_cpumap_attach test failed also. The test needs an 
adjustment in the .data_in and .data_size_in to have at least ETH_HLEN.

pw-bot: cr