mount(2) with NULL fstype returns ENOENT instead of EINVAL after introducing fs_context methods

mount(2) with NULL fstype returns ENOENT instead of EINVAL after introducing fs_context methods

[Murphy Zhou]

Hi,

Since this
commit f3a09c92018a91ad0981146a4ac59414f814d801 Author: Al Viro
<viro@zeniv.linux.org.uk>
Date:   Sun Dec 23 18:55:56 2018 -0500

    introduce fs_context methods

    Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>


LTP[1] mount02 tests #5 starts to fail.

This testcase is calling mount(2) with a  NULL fs type. It's expecting
EINVAL to PASS but now we are getting ENOENT.

Should I fix the testcase or there is some work in the kernel ? I'm
wring this email to confirm with you guys.

Thanks,
M
[1] https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/syscalls/mount/mount02.c

Re: mount(2) with NULL fstype returns ENOENT instead of EINVAL after introducing fs_context methods

[Karel Zak]

On Thu, Mar 14, 2019 at 03:20:05PM +0800, Murphy Zhou wrote:
> Hi,
> 
> Since this
> commit f3a09c92018a91ad0981146a4ac59414f814d801 Author: Al Viro
> <viro@zeniv.linux.org.uk>
> Date:   Sun Dec 23 18:55:56 2018 -0500
> 
>     introduce fs_context methods
> 
>     Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> 
> 
> LTP[1] mount02 tests #5 starts to fail.
> 
> This testcase is calling mount(2) with a  NULL fs type. It's expecting
> EINVAL to PASS but now we are getting ENOENT.
> 
> Should I fix the testcase or there is some work in the kernel ? I'm
> wring this email to confirm with you guys.

It's necessary to fix kernel, because userspace error messages are
generated by error codes. 

   EINVAL means message "wrong fs type, bad option, bad..."
   ENOENT means message "mount(2) system call failed: %m"

 Karel


-- 
 Karel Zak  <kzak@redhat.com>
 http://karelzak.blogspot.com

Re: mount(2) with NULL fstype returns ENOENT instead of EINVAL after introducing fs_context methods

[Murphy Zhou]

FIX: It's calling mount(2) with a NULL device name.

On Thu, Mar 14, 2019 at 3:20 PM Murphy Zhou <jencce.kernel@gmail.com> wrote:
>
> Hi,
>
> Since this
> commit f3a09c92018a91ad0981146a4ac59414f814d801 Author: Al Viro
> <viro@zeniv.linux.org.uk>
> Date:   Sun Dec 23 18:55:56 2018 -0500
>
>     introduce fs_context methods
>
>     Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
>
>
> LTP[1] mount02 tests #5 starts to fail.
>
> This testcase is calling mount(2) with a  NULL fs type. It's expecting
> EINVAL to PASS but now we are getting ENOENT.
>
> Should I fix the testcase or there is some work in the kernel ? I'm
> wring this email to confirm with you guys.
>
> Thanks,
> M
> [1] https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/syscalls/mount/mount02.c