[Bug python/20077] New: gdb.Value string object has strlen of 1 after backtrace

[kieranbingham at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Bug ID: 20077
Summary: gdb.Value string object has strlen of 1 after
backtrace
Product: gdb
Version: HEAD
Status: NEW
Severity: normal
Priority: P2
Component: python
Assignee: unassigned at sourceware dot org
Reporter: kieranbingham at gmail dot com
Target Milestone: ---

So this is an odd bug to reproduce.

As part of integrating extra debug commands into the Linux Kernel, we created
an lx-version command. It's simple enough and just finds, and displays the
version string from the kernel.


http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/scripts/gdb/linux/proc.py

However, if you run a backtrace, before you use the python lx-version command,
then the string returned prints only one character.
It actually has a strlen of 0 - but I traced through the gdb.Value code, and
found that somewhere it gets +1 added to it, which means it prints just one
char.


(gdb) bt
#0 cpu_v7_do_idle () at .../linux/arch/arm/mm/proc-v7.S:74
#1 0xc0308728 in arch_cpu_idle () at .../linux/arch/arm/kernel/process.c:72
#2 0xc0376b28 in cpuidle_idle_call () at .../linux/kernel/sched/idle.c:151
#3 cpu_idle_loop () at .../linux/kernel/sched/idle.c:242
#4 cpu_startup_entry (state=<optimized out>) at
.../linux/kernel/sched/idle.c:291
#5 0xc0ae8a30 in rest_init () at .../linux/init/main.c:408
#6 0xc0f00c5c in start_kernel () at .../linux/init/main.c:661
#7 0x8020807c in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
(gdb) lx-version
L(gdb) print linux_banner #### Note the 'L' at the beginning of the line
####
$1 = 0xc0b00074 <linux_banner> "Linux version 4.6.0-rc1
(***@CookieMonster) (gcc version 5.2.1 20151010 (Ubuntu 5.2.1-22ubuntu1) )
#13 SMP Thu Mar 31 10:33:19 BST 2016\n"
(gdb) print &linux_banner
$2 = (const char (*)[]) 0xc0b00074 <linux_banner>
(gdb) lx-version
L(gdb)

If however, upon connecting to the target, you run 'lx-version', *then* 'bt',
both operate as expected.
Perhaps the string / value gets cached somewhere?

/me is puzzled :D

[kieranbingham at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Kieran Bingham <kieranbingham at gmail dot com> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |jan.kiszka at siemens dot com,
| |kieranbingham at gmail dot com

[peter.griffin at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Peter Griffin <peter.griffin at linaro dot org> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |peter.griffin at linaro dot org

[palves at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Pedro Alves <palves at redhat dot com> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |palves at redhat dot com

--- Comment #1 from Pedro Alves <palves at redhat dot com> ---
Did you ever find out what's going on?

[kieranbingham at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

--- Comment #2 from Kieran Bingham <kieranbingham at gmail dot com> ---
Hi Pedro,

No we never did quite get to the bottom of this.

I'm afraid I'm on honeymoon now so apologies for the delays, I'll find you on
IRC when I'm back and we can see what we can do to get this resolved or at
least understand the root cause.

[peter.griffin at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

--- Comment #3 from Peter Griffin <peter.griffin at linaro dot org> ---
Hi,

I just entered this bug in the Linaro GDB bugzilla, in the hope that somebody
in the Linaro toolchain working group will have some time to debug this issue.

Unfortunately I haven't had time to debug it further, as I'm busy trying to
prepare the linux-kthread GDB patches to send to the ML.

Pete.

[tromey at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Tom Tromey <tromey at sourceware dot org> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |tromey at sourceware dot org

--- Comment #4 from Tom Tromey <tromey at sourceware dot org> ---
Does this still happen?

[tromey at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Tom Tromey <tromey at sourceware dot org> changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |WAITING

[kieranbingham at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

--- Comment #5 from Kieran Bingham <kieranbingham at gmail dot com> ---
Hi Tom,

It's a long time since I've tried this.

I'll try to see if I can recreate this issue this week.
Please ping me, either here or on IRC if it slips my radar.
--
Kieran

[kieranbingham at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

Kieran Bingham <kieranbingham at gmail dot com> changed:

What |Removed |Added
----------------------------------------------------------------------------
Version|HEAD |7.3

--- Comment #6 from Kieran Bingham <kieranbingham at gmail dot com> ---
So I've been sent a patch that simply adds a cast to the string pointer here,
to workaround this issue.

In the patch - the submitter describes this issue as only happening on GBD 7.3
and below.

https://lore.kernel.org/lkml/20181017023652.6867-1-***@gmail.com/t/#u

I'm afraid I haven't been able to find time to replicate this issue again yet.
But if it's not causing problems in mainline GDB then I don't think there's
much action to take.

We'll likely apply the cast patch to the linux kernel command to support older
versions if it works. It should be 'forwards' compatible.

[kieranbingham at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20077

--- Comment #7 from Kieran Bingham <kieranbingham at gmail dot com> ---
So this issue is still present in mainline GDB.

The following thread [0] shows the workaround which will handle this for us.
But the bug might still be interesting for someone to investigate:

[0] https://lkml.org/lkml/2018/10/17/798