Re: WARNING in do_jobctl_trap

[Dmitry Vyukov]

On Wed, Jan 27, 2016 at 5:04 PM, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> On 01/26, Oleg Nesterov wrote:
>>
>> On 01/26, Dmitry Vyukov wrote:
>> >
>> > Hello,
>> >
>> > The following program triggers WARN_ON_ONCE(!signr) in do_jobctl_trap:
>>
>> Can't reproduce, but most probably I understand the problem... I'll try to
>> make the simplified test-case tomorrow, it should work if I am right.
>
> This triggers the same problem immediately:
>
>
>         #include <unistd.h>
>         #include <sys/ptrace.h>
>         #include <sys/wait.h>
>
>         void test(void)
>         {
>                 for (;;) {
>                         if (fork()) {
>                                 wait(NULL);
>                                 continue;
>                         }
>
>                         ptrace(PTRACE_SEIZE, getppid(), 0, 0);
>                         ptrace(PTRACE_INTERRUPT, getppid(), 0, 0);
>                         _exit(0);
>                 }
>         }
>
>         int main(void)
>         {
>                 int np;
>
>                 for (np = 0; np < 8; ++np)
>                         if (!fork())
>                                 test();
>
>                 while (wait(NULL) > 0)
>                         ;
>                 return 0;
>         }
>
> and just in case, your test-case works too if I run
>
>         perl -e 'fork; fork; fork; 1 while 1' &
>
> in the background, it has too many delays to work on the idle system.
>
>> Does it work for you reliably?  If yes, any chance you can try the patch
>> below?
>>
>> Oleg.
>>
>> --- x/kernel/ptrace.c
>> +++ x/kernel/ptrace.c
>> @@ -73,11 +73,11 @@ void __ptrace_unlink(struct task_struct *child)
>>  {
>>       BUG_ON(!child->ptrace);
>>
>> -     child->ptrace = 0;
>>       child->parent = child->real_parent;
>>       list_del_init(&child->ptrace_entry);
>>
>>       spin_lock(&child->sighand->siglock);
>> +     child->ptrace = 0;
>>
>>       /*
>>        * Clear all pending traps and TRAPPING.  TRAPPING should be
>
> The change above helps, but I need to think more...


Hi Oleg,

I am running with the __ptrace_unlink fix since then, and did not see any bugs.
Please mail it a proper patch.