Hi Ulrich and community,

Please find attached the patch.

>Two issues:

>- Please only use current_inferior() in the ::mourn_inferior path.
>  In the ::detach path you still should use the passed-in "inf".
 > (This means pd_enable/disable still should get an "inf" argument.)

This is done. I misunderstood the previous email. Apologise for the same.

So now the inferior argument is passed in pd_enable () and pd_disable () so that we are good and do not cause the memory leak and pid != 0 assertion failure.. Outputs before and after patch remain same as the previous email.

Kindly check in this version.

Have a nice day ahead. Appreciate your patience.

Thanks and regards,
Aditya.


From: Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
Date: Tuesday, 2 May 2023 at 7:36 PM
To: gdb-patches@sourceware.org <gdb-patches@sourceware.org>, Aditya Kamath1 <Aditya.Kamath1@ibm.com>
Cc: Sangamesh Mallayya <sangamesh.swamy@in.ibm.com>
Subject: Re: [Patch] Fix Assertion pid 0 failure in AIX while running gdb.threads/foll-fork-other-thread
Aditya Kamath1 <Aditya.Kamath1@ibm.com> wrote:

>>Ah, right. There's indeed no inf argument (unfortunately), but you can
>>just use "current_inferior ()" instead.  This is guaranteed to be
>>set correctly at this point.
>
>This makes it easy. Thanks Ulrich. It fixes the issue and no memory leaks.
>Kindly push this small patch if there are no further changes.

Two issues:

- Please only use current_inferior() in the ::mourn_inferior path.
  In the ::detach path you still should use the passed-in "inf".
  (This means pd_enable/disable still should get an "inf" argument.)

- Your patch adds:
    data = get_thread_data_helper_for_pid (current_inferior ()->pid);
  where get_thread_data_helper_for_pid does:
    inferior *inf = find_inferior_pid (current_inferior ()->process_target (),
                                       pid);
    return get_aix_thread_variables_data (inf);
  This seems a bit silly (and inefficient) - if you already have
  an "inf", you should just use get_aix_thread_variables_data directly.

Bye,
Ulrich