Re: [PATCH bpf] bpf: Add extra path pointer check to d_path helper

[Stanislav Fomichev <sdf@google.com>]

On 06/04, Jiri Olsa wrote:
> Anastasios reported crash on stable 5.15 kernel with following
> bpf attached to lsm hook:
> 
>   SEC("lsm.s/bprm_creds_for_exec")
>   int BPF_PROG(bprm_creds_for_exec, struct linux_binprm *bprm)
>   {
>           struct path *path = &bprm->executable->f_path;
>           char p[128] = { 0 };
> 
>           bpf_d_path(path, p, 128);
>           return 0;
>   }
> 
> but bprm->executable can be NULL, so bpf_d_path call will crash:
> 
>   BUG: kernel NULL pointer dereference, address: 0000000000000018
>   #PF: supervisor read access in kernel mode
>   #PF: error_code(0x0000) - not-present page
>   PGD 0 P4D 0
>   Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC NOPTI
>   ...
>   RIP: 0010:d_path+0x22/0x280
>   ...
>   Call Trace:
>    <TASK>
>    bpf_d_path+0x21/0x60
>    bpf_prog_db9cf176e84498d9_bprm_creds_for_exec+0x94/0x99
>    bpf_trampoline_6442506293_0+0x55/0x1000
>    bpf_lsm_bprm_creds_for_exec+0x5/0x10
>    security_bprm_creds_for_exec+0x29/0x40
>    bprm_execve+0x1c1/0x900
>    do_execveat_common.isra.0+0x1af/0x260
>    __x64_sys_execve+0x32/0x40
> 
> It's problem for all stable trees with bpf_d_path helper, which was
> added in 5.9.
> 
> This issue is fixed in current bpf code, where we identify and mark
> trusted pointers, so the above code would fail to load.
> 
> For the sake of the stable trees and to workaround potentially broken
> verifier in the future, adding the code that reads the path object from
> the passed pointer and verifies it's valid in kernel space.
> 
> Cc: stable@vger.kernel.org # v5.9+
> Fixes: 6e22ab9da793 ("bpf: Add d_path helper")
> Suggested-by: Alexei Starovoitov <ast@kernel.org>
> Reported-by: Anastasios Papagiannis <tasos.papagiannnis@gmail.com>
> Signed-off-by: Jiri Olsa <jolsa@kernel.org>

Acked-by: Stanislav Fomichev <sdf@google.com>

One question though: does it really have to go via bpf tree? Can it
be a stable-only fix? Otherwise it's not really clear why we
need to double-check anything if the pointer is trusted..

> ---
>  kernel/trace/bpf_trace.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
> index 9a050e36dc6c..aecd98ee73dc 100644
> --- a/kernel/trace/bpf_trace.c
> +++ b/kernel/trace/bpf_trace.c
> @@ -900,12 +900,22 @@ static const struct bpf_func_proto bpf_send_signal_thread_proto = {
>  
>  BPF_CALL_3(bpf_d_path, struct path *, path, char *, buf, u32, sz)
>  {
> +	struct path copy;
>  	long len;
>  	char *p;
>  
>  	if (!sz)
>  		return 0;
>  
> +	/*
> +	 * The path pointer is verified as trusted and safe to use,
> +	 * but let's double check it's valid anyway to workaround
> +	 * potentially broken verifier.
> +	 */
> +	len = copy_from_kernel_nofault(&copy, path, sizeof(*path));
> +	if (len < 0)
> +		return len;
> +
>  	p = d_path(path, buf, sz);
>  	if (IS_ERR(p)) {
>  		len = PTR_ERR(p);
> -- 
> 2.40.1
>