Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <dcb534d5-e9b7-74b3-4088-eb498adb9ec2@infradead.org>
Date:   Tue, 3 May 2022 10:26:04 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.8.1
Subject: Re: BUG: vfork() returns EINVAL after unshare(CLONE_NEWTIME)
Content-Language: en-US
To:     =?UTF-8?B?0JzQsNGA0Log0JrQvtGA0LXQvdCx0LXRgNCz?= 
        <socketpair@gmail.com>, linux-kernel@vger.kernel.org,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Christian Brauner <christian@brauner.io>
References: <CAEmTpZGJxtTuq8LDTLDK4V8to_Z30WEmo=p_2umboHvZaDr=Xg@mail.gmail.com>
From:   Randy Dunlap <rdunlap@infradead.org>
In-Reply-To: <CAEmTpZGJxtTuq8LDTLDK4V8to_Z30WEmo=p_2umboHvZaDr=Xg@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

[add some readers:]

On 3/31/22 01:36, Марк Коренберг wrote:
> test:
> ----------------------------------------
> #define _GNU_SOURCE 1
> #include <stdio.h>
> #include <sched.h>
> #include <stdlib.h>
> #include <unistd.h>
> #include <sys/types.h>
> #include <sys/wait.h>
> #include <err.h>
> 
> #ifndef CLONE_NEWTIME
> #define CLONE_NEWTIME   0x00000080
> #endif
> 
> int main (void)
> {
>   if (unshare (CLONE_NEWTIME))  err (EXIT_FAILURE, "UNSHARE_NEWTIME");
> 
>   pid_t pid;
>   switch (pid=vfork ())
>   {
>   case 0:
>     _exit(0);
>   case -1:
>     err(EXIT_FAILURE, "vfork BUG");
>   default:
>     waitpid(pid, NULL, 0);
>   }
>   return 0;
> }
> -----------------------------------------------
> 
> Seems the bug in kernel/fork.c. It does not check for CLONE_VFORK
> 
> if (clone_flags & (CLONE_THREAD | CLONE_VM)) {
>     if (nsp->time_ns != nsp->time_ns_for_children)
>         return ERR_PTR(-EINVAL);
> }
> 
> -----------------------------------
> Changing vfork() to fork() in the program works ok.
> The bug has been found during regression in our system because Python
> 3.10 runs subprocesses using vfork()+execve().
> 

-- 
~Randy