Bug 35881 - Disassembly listing additions(?)
Summary: Disassembly listing additions(?)
Status: NEW
Alias: None
Product: CDT
Classification: Tools
Component: cdt-debug (show other bugs)
Version: 2.0   Edit
Hardware: All All
: P3 enhancement with 1 vote (vote)
Target Milestone: ---   Edit
Assignee: cdt-debug-inbox@eclipse.org CLA
QA Contact: Jonah Graham CLA
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-03-31 14:46 EST by Johan CLA
Modified: 2020-09-04 15:23 EDT (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Johan CLA 2003-03-31 14:46:50 EST
Hi,

After some trouble, I am slowly getting the CDT debugger
to work with an embedded CPU simulator debugger 
(arm-elf-gdb.exe in my case).

After successully loading an application.elf file into the
debugger, and switching into disassembler mode, I get
an ARM disassembly listing similar to:

0x00000660 <main+20>:  mov\tr2, #2\t; 0x2
0x00000664 <main+24>:  mvn\tr3, #48896\t; 0xbf00
0x00000668 <main+28>:  sub\tr3, r3, #223\t; 0xdf
0x0000066c <main+32>:  str\tr2, [r3]
0x00000670 <main+36>:  mvn\tr2, #48896\t; 0xbf00
0x00000674 <main+40>:  ldr\tr3, [r2, -#207]

As it seems, the  \t  codes should preferably
be expanded into a tab for proper column alignment.
Can this be added?

Furthermore, I would like to see the hex opcodes
for every assembler instruction, side-by-side with
the address. Could this be added as well?

Kind regards,
Johan Johansson
Comment 1 Nobody - feel free to take it CLA 2003-03-31 15:04:06 EST
The '\t' problem has already fixed in the head branch by Chris Songer.
Regarding your second request. Do you know if there is a way to retrieve the 
hexadecimal opcode from gdb?
Comment 2 Johan CLA 2003-04-01 11:23:41 EST
Unfortunately I don't know very much about gdb, but
if you know the lenght and address of every assembler 
instruction, it should really be a matter of reading
the memory contents for X number of bytes starting
at address Y for every instruction, I assume?

(The opcodes is just a hex view of the memory 
bytes occupied by the instruction)

/Johan
Comment 3 Nobody - feel free to take it CLA 2003-04-01 11:49:05 EST
Yes, it is possible to read memory for each opcode but it is very time 
consuming. What if we use hovering instead?
Comment 4 Johan CLA 2003-04-02 12:19:52 EST
OK,

I think that the default behaviour in embedded debuggers really are to show 
the address, opcode and disasm string in parallel columns, but if it is very 
slow perhaps that is not viable. 

Perhaps a GUI choice "show opcodes on/off" could work, or even hovering 
tooltips as mentioned. Anyone else having any preferences on this?

/Johan
Comment 5 Nobody - feel free to take it CLA 2003-04-02 12:49:57 EST
You are the first who raised this question.
I agree that the best solution is to show opcode values in the view and to have 
an "on/off" preference for it. I'll try it to see how it slows down the 
debugging. 
Comment 6 Kleo Hapitas CLA 2004-07-07 16:54:05 EDT
PR was not targeted to any particular release. Changing target milestone to 2.1
Comment 7 Nobody - feel free to take it CLA 2004-11-15 11:25:18 EST
Moving the target milestone to 3.0.
Comment 8 Nobody - feel free to take it CLA 2005-06-10 11:32:42 EDT
Deferred.
Comment 9 Nobody - feel free to take it CLA 2005-09-09 14:42:10 EDT
Reassigning to the pool.
Comment 10 Doug Schaefer CLA 2007-08-21 11:01:15 EDT
Future means you commit to fix it in the Future. Inboxes can't make committments. Moving to '--'.