On 2015/5/29 18:23, Jiri Jaburek wrote:
> On 05/28/15 14:33, Jan Stancek wrote:
>> ----- Original Message -----
>>> From: "Yuan Sun" <[email protected]>
>>> To: [email protected]
>>> Cc: [email protected], [email protected],
>>> [email protected]
>>> Sent: Wednesday, 27 May, 2015 11:00:21 PM
>>> Subject: [PATCH V2] containers: new testcase userns02
>>>
>>> The user ID and group ID, which are inside a container, can
>>> be modified by its parent process.
>>>
>>> Signed-off-by: Yuan Sun <[email protected]>
>> Hi,
>>
>> couple comments inline, but overall it looks good to me.
>> Unless someone points out other issues, I can fix these before commit.
>>
> As a FYI to the author - there's a "libclone" library in containers/
> that can take care of the forking/cloning/unsharing, specifically the
> do_clone_tests function could be used (uses ltp_clone_quick).
> Given the simplicity of the function, though, I wouldn't see it as
> a major issue.
Hi Jan,
Thanks for your comment. In general, I agree with you. But in this
case I have different opinions.
I tried replacing ltp_clone_quick with do_clone_tests.
#childpid = do_clone_tests(CLONE_NEWUSER | SIGCHLD, (void *)child_fn1,
NULL, NULL, NULL);
However, I found "do_clone_tests" function can't return the child
pid. In this case,
the parent needs the child pid to modify /proc/$childpid/uid_map. Maybe
ltp_clone_quick
is more appropriate than do_clone_tests in this case.
Of course, if you have different opinion please feel free to let me know.
Regards.
Yuan
>
>
> .
>
------------------------------------------------------------------------------
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
