This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: further help on dwarf2 debug

From: Richard Henderson <rth at redhat dot com>
To: IainS <developer at sandoe-acoustics dot co dot uk>
Cc: GCC Development <gcc at gcc dot gnu dot org>
Date: Thu, 24 Sep 2009 09:41:57 -0700
Subject: Re: further help on dwarf2 debug
References: <93F69288-1096-4ED3-8D5C-F186479CF617@sandoe-acoustics.co.uk>

On 09/24/2009 12:35 AM, IainS wrote:

The main difference (and the area causing the tool to barf) is that the
FDE on 4.5 has some extra information not present in 4.4

.byte 0x4 # DW_CFA_advance_loc4
.set L$set$6,LCFI3-LCFI1
.long L$set$6
.byte 0xc5 # DW_CFA_restore, column 0x5
.byte 0xc # DW_CFA_def_cfa
.byte 0x4 # uleb128 0x4
.byte 0x4 # uleb128 0x4

...

What is the purpose of the addition?

It describes the operation of the "leave" instruction.

If you set a breakpoint on the "ret" instruction, you now have correct unwind information at that spot. The debugger does not have to resort to code inspection to figure out the backtrace.

While gdb and most likely other more robust user-level debuggers *will* resort to code inspection (because historically it was necessary), newer tools like SystemTap do not; SystemTap is written to work only with the debug information.

r~

References:
- further help on dwarf2 debug
  - From: IainS

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]