Re: [PATCH bpf v2 3/3] selftest/bpf: Test the read of vsyscall page under x86-64

[Yonghong Song <yonghong.song@linux.dev>]

On 1/26/24 3:54 AM, Hou Tao wrote:
> From: Hou Tao <houtao1@huawei.com>
>
> Under x86-64, when using bpf_probe_read_kernel{_str}() or
> bpf_probe_read{_str}() to read vsyscall page, the read may trigger oops,
> so add one test case to ensure that the problem is fixed. Beside those
> four bpf helpers mentioned above, testing the read of vsyscall page by
> using bpf_probe_read_user{_str} and bpf_copy_from_user{_task}() as well.
>
> The test case passes the address of vsyscall page to these six helpers
> and checks whether the returned values are expected:
>
> 1) For bpf_probe_read_kernel{_str}()/bpf_probe_read{_str}(), the
>     expected return value is -ERANGE as shown below:
>
> bpf_probe_read_kernel_common
>    copy_from_kernel_nofault
>      // false, return -ERANGE
>      copy_from_kernel_nofault_allowed
>
> 2) For bpf_probe_read_user{_str}(), the expected return value is -EFAULT
>     as show below:
>
> bpf_probe_read_user_common
>    copy_from_user_nofault
>      // false, return -EFAULT
>      __access_ok
>
> 3) For bpf_copy_from_user(), the expected return value is -EFAULT:
>
> // return -EFAULT
> bpf_copy_from_user
>    copy_from_user
>      _copy_from_user
>        // return false
>        access_ok
>
> 4) For bpf_copy_from_user_task(), the expected return value is -EFAULT:
>
> // return -EFAULT
> bpf_copy_from_user_task
>    access_process_vm
>      // return 0
>      vma_lookup()
>      // return 0
>      expand_stack()
>
> The occurrence of oops depends on the availability of CPU SMAP [1]
> feature and there are three possible configurations of vsyscall page in
> boot cmd-line: vsyscall={xonly|none|emulate}, so there are totally six
> possible combinations. Under all these combinations, the running of the
> test case succeeds.
>
> [1]: https://en.wikipedia.org/wiki/Supervisor_Mode_Access_Prevention
>
> Signed-off-by: Hou Tao <houtao1@huawei.com>

The first two patches look good to me but I think it would be better
if x86 folks can ack on them. The selftest patch LGTM.

Acked-by: Yonghong Song <yonghong.song@linux.dev>