[Bug server/16255] New: gdbserver cannot attach to a second inferior that is multi-threaded

[Bug server/16255] New: gdbserver cannot attach to a second inferior that is multi-threaded

[marc.khouzam at ericsson dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16255

            Bug ID: 16255
           Summary: gdbserver cannot attach to a second inferior that is
                    multi-threaded
           Product: gdb
           Version: 7.4
            Status: NEW
          Severity: normal
          Priority: P2
         Component: server
          Assignee: unassigned at sourceware dot org
          Reporter: marc.khouzam at ericsson dot com

This problem was not happening in version 7.3, but starts with 7.4.  I'm
running on Ubuntu 11.10 but we've seen it on Ubuntu 12.04 and even on a very
old SLED 10 machine.

When debugging with gdbserver, if I try to attach to a second process and that
process is multi-threaded, gdbserver will fail to attach to any thread beyond
the first one with repeating error:

Cannot attach to lwp 13779: Operation not permitted (1)
Cannot attach to lwp 13779: Operation not permitted (1)
Cannot attach to lwp 13779: Operation not permitted (1)
Cannot attach to lwp 13779: Operation not permitted (1)
Cannot attach to lwp 13779: Operation not permitted (1)
Cannot attach to lwp 13779: Operation not permitted (1)
Cannot attach to lwp 13779: Operation not permitted (1)
[forever and very fast]

The problem only happens with gdbserver, local attach works ok.
The problem only happens when attaching, starting a second new multi-threaded
process from GDB works ok.
Attaching to a second process, if that process in single-threaded, does work. 
In fact, attaching to a second process with a single thread works, even if that
process then creates new threads after the attach.

To reproduce:
1- start two instances of a long running multi-threaded binary from a shell
2- start gdbserver: gdbserver --multi :9999
4- debug with gdb:

> gdb.7.7 /home/lmckhou/testing/multithread
GNU gdb (GDB) 7.6.50.20131126-cvs
Reading symbols from /home/lmckhou/testing/multithread...done.
(gdb) l
3
4       void *thread_exec(void *p) {
5           for (int i=0;i<500;i++) {
6               sleep(1);
7           }
8       }
9
10      int main() {
11          pthread_t thread2;
12          int iret2 = pthread_create(&thread2, NULL, thread_exec, (void*)
"Thread 2");
13          pthread_join(thread2, NULL);
14          return 0;
15      }
16
(gdb) set target-async on
(gdb) set non-stop on
(gdb) set pagination off
(gdb) add-inferior -exec /home/lmckhou/testing/multithread
Added inferior 2
Reading symbols from /home/lmckhou/testing/multithread...done.
(gdb) tar ext :9999
Remote debugging using :9999
(gdb) attach 13896
Attaching to program: /home/lmckhou/testing/multithread, process 13896
[New Thread 13896.13896]
[New Thread 13896.13898]
(gdb) inferior 2
[Switching to inferior 2 [<null>] (/home/lmckhou/testing/multithread)]

=== The first inferior could have been single threaded too ===
=== Now we will attach to another multi-threaded process ===

(gdb) attach 13897
Attaching to program: /home/lmckhou/testing/multithread, process 13897

=== here gdbserver goes crazy printing the error message: ===
=== "Cannot attach to lwp 13779: Operation not permitted (1)" ===

=== after some delay, we see gdb printing: ===

Ignoring packet error, continuing...
Attaching to process 13897 failed with: timeout

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug server/16255] gdbserver cannot attach to a second inferior that is multi-threaded

[simon.marchi at polymtl dot ca]

https://sourceware.org/bugzilla/show_bug.cgi?id=16255

Simon Marchi <simon.marchi at polymtl dot ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |simon.marchi at polymtl dot ca

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug server/16255] gdbserver cannot attach to a second inferior that is multi-threaded

[simon.marchi at polymtl dot ca]

https://sourceware.org/bugzilla/show_bug.cgi?id=16255

--- Comment #1 from Simon Marchi <simon.marchi at polymtl dot ca> ---
Ok, the problem is that a ptid with a wrong value is built when we attach the
subsequent threads.

See:
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=blob;f=gdb/gdbserver/linux-low.c;hb=HEAD#l673

For the pid field, we use the pid of current_inferior. When attaching to the
second inferior, current_inferior is never set to the newly created inferior,
it still points to the first one. Therefore, ptids of the threads of the second
inferior are created with the pid of the first.

One solution is to temporarily switch current_inferior in linux_attach when
attaching the subsequent threads.

Another would be to change the line
    ptid = ptid_build (pid, lwpid, 0);
for
    ptid = ptid_build (linux_proc_get_tgid (lwpid), lwpid, 0);

which will go read the pid (tgid) in /proc.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug server/16255] gdbserver cannot attach to a second inferior that is multi-threaded

[cvs-commit at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=16255

--- Comment #3 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> ---
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "gdb and binutils".

The branch, master has been updated
       via  94611da2409673314d7caaf38c66baa52d6f5adc (commit)
      from  7ae1a6a6ccda41aa8bbe9adb0f7fcde8bf8d5cb3 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=94611da2409673314d7caaf38c66baa52d6f5adc

commit 94611da2409673314d7caaf38c66baa52d6f5adc
Author: Pedro Alves <palves@redhat.com>
Date:   Fri Apr 25 19:22:27 2014 +0100

    Add missing gdbserver/ChangeLog entry for previous commit.

    2014-04-25  Pedro Alves  <palves@redhat.com>

        PR server/16255
        * linux-low.c (linux_attach_fail_reason_string): New function.
        (linux_attach_lwp): Delete.
        (linux_attach_lwp_1): Rename to ...
        (linux_attach_lwp): ... this.  Take a ptid instead of a pid as
        argument.  Remove "initial" parameter.  Return int instead of
        void.  Don't error or warn here.
        (linux_attach): Adjust to call linux_attach_lwp.  Call error on
        failure to attach to the tgid.  Call warning when failing to
        attach to an lwp.
        * linux-low.h (linux_attach_lwp): Take a ptid instead of a pid as
        argument.  Remove "initial" parameter.  Return int instead of
        void.  Don't error or warn here.
        (linux_attach_fail_reason_string): New declaration.
        * thread-db.c (attach_thread): Adjust to linux_attach_lwp's
        interface change.  Use linux_attach_fail_reason_string.

-----------------------------------------------------------------------

Summary of changes:
 gdb/gdbserver/ChangeLog |   19 +++++++++++++++++++
 1 files changed, 19 insertions(+), 0 deletions(-)

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug server/16255] gdbserver cannot attach to a second inferior that is multi-threaded

[palves at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=16255

Pedro Alves <palves at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|---                         |7.8

--- Comment #4 from Pedro Alves <palves at redhat dot com> ---
Fixed.

-- 
You are receiving this mail because:
You are on the CC list for the bug.