public inbox for gdb-prs@sourceware.org
help / color / mirror / Atom feed
From: "tom.k.cook at gmail dot com" <sourceware-bugzilla@sourceware.org>
To: gdb-prs@sourceware.org
Subject: [Bug fortran/17989] New: gdb hangs on some Fortran stack frames
Date: Tue, 17 Feb 2015 05:46:00 -0000	[thread overview]
Message-ID: <bug-17989-4717@http.sourceware.org/bugzilla/> (raw)

https://sourceware.org/bugzilla/show_bug.cgi?id=17989

            Bug ID: 17989
           Summary: gdb hangs on some Fortran stack frames
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: fortran
          Assignee: unassigned at sourceware dot org
          Reporter: tom.k.cook at gmail dot com

Overview:  GDB hangs when trying to display some Fortran stack frames.

Platform: Linux 3.13.0 x86_64 SMP (Ubuntu Trusty)

Steps to reproduce:  Unknown.  Sorry!  I'll give as much information as I can.

The software being debugged is mixed Fortran/C++.  The main method is in
Fortran but the executable is linked using G++.  The compiler is GCC 4.8 (for
both Fortran and C++) but I have confirmed the bug when compiling with GCC 4.9.
 I have a fairly recent GCC HEAD (early January) build around and could try
that if it might be useful.

GDB hangs when attempting to display certain stack frames (eg. when a
breakpoint is hit, or when 'finish' is used to move up a frame, or when 'bt' is
used).  'info line' and 'list' still work in these situations.

The problem always happens with Fortran stack frames, including when there are
no C++ frames on the stack.

The Fortran code is compiled with '-g -cpp -ffixed-line-length-none
-ffree-line-length-none -fcoarray=single -fno-underscoring -fPIC'.

The executable is linked with '-g', objects and libraries.

When gdb hangs, it is consuming 100% CPU (ie one CPU core).

The hang can be interrupted with Ctrl-C and the debugging session continued
normally (so long as you don't want to display the call stack etc).

I have checked that both GDB 7.7 (Ubuntu Trusty) and a fresh build from git
latest today, using GCC-5.0 latest-as-of-early-January, display this problem.

I have tried to reproduce the bug on a very simple Fortran/C++ project but have
not been able to.

Any suggestions on how to fix this, or on steps I might take to reliably
reproduce it, will be gratefully received.  I have a build of gdb
latest-of-today and could debug gdb itself if this will be useful. 
Unfortunately the Fortran/C++ code being debugged when the problem occurs is
commercial and can't be shared.

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


             reply	other threads:[~2015-02-17  4:46 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-02-17  5:46 tom.k.cook at gmail dot com [this message]
2015-02-17  6:25 ` [Bug fortran/17989] " tom.k.cook at gmail dot com
2015-02-17  9:45 ` tom.k.cook at gmail dot com
2015-02-20 12:08 ` tom.k.cook at gmail dot com

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=bug-17989-4717@http.sourceware.org/bugzilla/ \
    --to=sourceware-bugzilla@sourceware.org \
    --cc=gdb-prs@sourceware.org \
    /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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).