Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Pedro Alves <palves@redhat.com>
To: ali_anwar <ali_anwar@codesourcery.com>
Cc: Tom Tromey <tromey@redhat.com>, gdb-patches@sourceware.org
Subject: Re: Updated patch for Bug 13217 - thread apply all detach throws a SEGFAULT
Date: Tue, 11 Dec 2012 16:43:00 -0000	[thread overview]
Message-ID: <50C7628F.5080004@redhat.com> (raw)
In-Reply-To: <50C75319.9080903@codesourcery.com>

On 12/11/2012 03:36 PM, ali_anwar wrote:
> +  if (thread_count ())
> +    {
> +      struct thread_info *tp_array;
> +      struct thread_info *tp;
> +      int i, k;
> +
> +      /* Save a copy of the thread_list in case we execute detach
> +         command.  */
> +      tp_array =  xmalloc (sizeof (struct thread_info) * thread_count ());

No need to compute the thread count twice, you can cache it.  No need to
copy the whole thread structure.  Make this an array of a thread
pointers, and then,

> +      for (i = 0, tp = thread_list; tp; i++, tp = tp->next)
> +        tp_array[i] = *tp;

      ALL_THREADS (tp)
        {
          tp_array[i] = tp;
          tp->refcount++;
        }

This increments the refcount of each current thread, so that attempts to
delete it just mark it as deleted (so the C object remains valid).

> +
> +      for (k = 0; k != i; k++)
> +        if (thread_alive (&tp_array[k]))

and then write:

      for (k = 0; k != i; k++)
        {
          if (thread_alive (tp_array[k]))
            {
              switch_to_thread (tp_array[k]->ptid);

              printf_filtered (_("\nThread %d (%s):\n"),
                               (tp_array->num, target_pid_to_str (inferior_ptid));
              execute_command (cmd, from_tty);
              strcpy (cmd, saved_cmd);        /* Restore exact command used
                                                 previously.  */
            }
        }

And put this in a cleanup:

      for (k = 0; k != i; k++)
        tp_array[k]->refcount--;

So that if the command throws an error, we still leave with the correct
refcounts.

The advantages are:

 - less memory necessary for the array.
 - handles the corner case of the target reusing a ptid (see
   add_thread_silent).  IOW, this way, even if the command happens to
   make the target reuse a ptid, "thread apply all" won't run the command
   on that new threads my mistake.



> +          {
> +            switch_to_thread ((&tp_array[k])->ptid);
> +
> +            printf_filtered (_("\nThread %d (%s):\n"),
> +                             (&tp_array[k])->num, target_pid_to_str (inferior_ptid));
> +            execute_command (cmd, from_tty);
> +            strcpy (cmd, saved_cmd);        /* Restore exact command used
> +                                               previously.  */
> +           }
> +       xfree (tp_array);

This xfree should be on the cleanup as well.

> +    }
>    do_cleanups (old_chain);

-- 
Pedro Alves


  reply	other threads:[~2012-12-11 16:43 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-09-26 11:46 ali_anwar
2012-09-26 11:50 ` ali_anwar
2012-09-27 15:27 ` Jan Kratochvil
2012-09-27 15:32 ` Tom Tromey
2012-12-10 18:37   ` ali_anwar
2012-12-10 20:20     ` Tom Tromey
2012-12-11 15:37       ` ali_anwar
2012-12-11 16:43         ` Pedro Alves [this message]
2013-07-10 10:31           ` ali_anwar
2013-07-10 12:57             ` Joel Brobecker
2013-07-10 17:55             ` Tom Tromey
2013-07-11 12:30               ` ali_anwar
2013-07-13 21:13                 ` Tom Tromey
2013-07-15 11:20                   ` ali_anwar

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=50C7628F.5080004@redhat.com \
    --to=palves@redhat.com \
    --cc=ali_anwar@codesourcery.com \
    --cc=gdb-patches@sourceware.org \
    --cc=tromey@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox