gdb without gcc - linux

Is it possible to run GDB with a program assembled with as and linked with ld? With gcc adding the flag -g allows for debugging but I get the error No symbol table is loaded. Use the "file" command when I try to add breakpoints to a loaded program.
Thanks!
EDIT Maybe I should make it clear that I'm learning and programming in assembly. All I really want is a stack trace but being able to use GDB would be great.
Resolution Running as -g does the trick.
Thank you to all that answered!!

It is possible. However, you need symbols in order to add symbolic breakpoints, and symbols are provided by debugging info; make sure your assembler and linker are providing those. EDIT With GNU as, use as -g. Or just use gcc -g: if you give it a .s file, it will invoke the assembler and linker as appropriate.
GDB understands debugging info in several formats: stabs, COFF, PE, DWARF, SOM. (Some of these are executable formats with debugging sections, others are debug info formats that can be embedded into executables, such as ELF.) gcc -g usually chooses whatever the platform's default is, gcc -ggdb usually chooses the most expressive (depending on your versions, possibly DWARF-3).
If you have debugging info embedded into or linked to by the executable, gdb will try to load it automatically. If you have it elsewhere, you may need to use file to tell gdb where to find it.
You can still debug without symbolic information. For example, you can issue break *0x89abcdef to insert a breakpoint at that address, if there's any code there.

you could try running as with the --gdwarf-2 or -g options, and make sure ld is not called with --strip-debug, and that your makefile/install process is not stripping the executable.

That's not an error preventing debugging, that's an error setting breakpoints in the way you are trying to do it. Since GDB doesn't have any symbol information, you'll have to set the breakpoints some other way.

If you don't have a symbol table, then you can't set breakpoints symbolically (by function name, line of code, etc). You could still set a breakpoint for a specific address, if you know the address you are trying to stop at.
gdb> b 0x12345678
Of course that's only useful if you know that you want to stop at 0x12345678
What does file say about your executable?

Related

Could not load shared library symbols for linux-vdso.so.1. while debugging

Not loading VDSO.so is one of the famous bugs you encounter while using gdb and glibc >2.2.
I found that was planned to get repaired in gdb 7.5.1, but it wasn't.
Okay I found a work-around here Here, but I didn't understand it so how to apply it.
OS: Arch Linux
IDE : QT creator 3.0.82
Compiler : GCC 4.8.2
NB: I am not sure if I am breaking the rules including the link above
Not loading VDSO.so is one of the famous bugs you encounter while using gdb and glibc >2.2.
No, it's not. The problem here is simply a useless warning, which you can safely ignore.
I found a work-around here Here, but I didn't understand it so how to apply it.
You didn't find a "workaround". You found a patch to GDB, which disables the warning.
To apply it, use the patch command, and then build your own GDB. But it is much simpler to just ignore the warning in the first place.
For anyone who (like me) just wants gdb to shut up about missing symbols, try adding this to your ~/.gdbinit (but see caveats below):
set logging redirect on
set logging file /dev/null
python
def on_new_objfile(e):
gdb.execute("set logging off")
#print "new objfile:",e.new_objfile.filename
if e.new_objfile.filename[:19] == "system-supplied DSO":
gdb.execute("set logging on") # hide inevitable error message
gdb.events.new_objfile.connect(on_new_objfile)
end
Caveats:
Monopolizes the set logging interface; if you want to use logging you'll need to change it to save the previous logging settings.
Hard-codes "system-supplied DSO"; might be brittle wrt new kernel or gdb versions.
It assumes at least one objfile will be loaded after vdso to reenable output; I'd be very interested if anyone with better gdb internals knowledge could point out the actual after-symbol-load-has-failed hook, since for now it risks leaving output disabled when the program starts if vdso is the last objfile loaded.

how to add linker flags to create compiler specific executable in scons

i am using scons for different compilers.
vc10 and renesas compiler.
if i compile the program by using env.program(---) am getting the link flags as
"link /nologo /subsystem:console /pdb:project.pdb /OUT:program.exe D:\build1\subdirA\subdirA.lib D:\build1\subdirB\subdirB.lib main.obj"
it is working for VC10 compiler. But for renesas(my microcontroller) compiler, i am getting an error like
"Cannot open file : "/OUT:program.exe""
it will accept " -output=program.abs" command when linking. how can i change that one. can you please tell me
when program is linking /OUT:program.exe is adding by default.
can you please tell me how to change that to "-output=program.abs"
Thank you
What you want to do is called cross-compile: compile a Renesas binary on Windows. Seems like what you have done is loaded the Windows VC10 toolset in SCons (SCons does this automatically unless told not to) and just changed the compiler binary, so SCons is still using the VC10 compiler/linker flags, which dont seem to be compatible. I had to do something similar once with SCons, where I cross-compiled Cavium Octeon in a Linux environment, but luckily for me almost all of the flags were compatible.
I dont know anything about Renesas, but if its compilation flags are more similar to another platform/toolset, then load those instead of Windows, as follows where Im loading the Linux gcc toolset.
env = Environment(tools = ['gcc'])
Look for Construction Environments in the SCons man page for a complete list of supported tools. Keep in mind that by doing this, you will no longer have support for the native platform toolset, Windows VC10 in your case.
When and if you find a similar platform, and you still need to change some compiler/linker flags or options, take a look at changing the related SCons Construction Variables. A few that could be helpful are: CXXFLAGS, LIBSUFFIX, LINKFLAGS, OBJSUFFIX, and PROGSUFFIX. The LINKFLAGS construction variable is actually the answer to your original question.
I did a google search for scons renesas, and came across this link which might also be helpful.

Linux, GNU GCC, ld, version scripts and the ELF binary format -- How does it work?

I'm trying to learn more about library versioning in Linux and how to put it all to work. Here's the context:
-- I have two versions of a dynamic library which expose the same set of interfaces, say libsome1.so and libsome2.so.
-- An application is linked against libsome1.so.
-- This application uses libdl.so to dynamically load another module, say libmagic.so.
-- Now libmagic.so is linked against libsome2.so. Obviously, without using linker scripts to hide symbols in libmagic.so, at run-time all calls to interfaces in libsome2.so are resolved to libsome1.so. This can be confirmed by checking the value returned by libVersion() against the value of the macro LIB_VERSION.
-- So I try next to compile and link libmagic.so with a linker script which hides all symbols except 3 which are defined in libmagic.so and are exported by it. This works... Or at least libVersion() and LIB_VERSION values match (and it reports version 2 not 1).
-- However, when some data structures are serialized to disk, I noticed some corruption. In the application's directory if I delete libsome1.so and create a soft link in its place to point to libsome2.so, everything works as expected and the same corruption does not happen.
I can't help but think that this may be caused due to some conflict in the run-time linker's resolution of symbols. I've tried many things, like trying to link libsome2.so so that all symbols are alised to symbol##VER_2 (which I am still confused about because the command nm -CD libsome2.so still lists symbols as symbol and not symbol##VER_2)... Nothing seems to work!!! Help!!!!!!
Edit: I should have mentioned it earlier, but the app in question is Firefox, and libsome1.so is libsqlite3.so shipped with it. I don't quite have the option of recompiling them. Also, using version scripts to hide symbols seems to be the only solution right now. So what really happens when symbols are hidden? Do they become 'local' to the SO? Does rtld have no knowledge of their existence? What happens when an exported function refers to a hidden symbol?
Try compiling both libsome1.so and libsome2.so to add symbol versioning, each with their own version (use the --version-script option to ld). Then link the application and libmagic.so using the new libraries. Then, libsome1.so and libsome2.so should be completely separate.
Problems can still occur if there are unversioned references to symbols. Such references can be satisfied by versioned definitions (so that it is possible to add symbol versioning to a library without breaking binary compatibility). If there are multiple symbols of the same name, it can sometimes be hard to predict which one will be used.
Regarding tools, nm -D does not display any information about symbol versioning. Try objdump -T or readelf -s instead.

Linux, GNU GCC, ld, version scripts and the ELF binary format -- How does it work? [duplicate]

I'm trying to learn more about library versioning in Linux and how to put it all to work. Here's the context:
-- I have two versions of a dynamic library which expose the same set of interfaces, say libsome1.so and libsome2.so.
-- An application is linked against libsome1.so.
-- This application uses libdl.so to dynamically load another module, say libmagic.so.
-- Now libmagic.so is linked against libsome2.so. Obviously, without using linker scripts to hide symbols in libmagic.so, at run-time all calls to interfaces in libsome2.so are resolved to libsome1.so. This can be confirmed by checking the value returned by libVersion() against the value of the macro LIB_VERSION.
-- So I try next to compile and link libmagic.so with a linker script which hides all symbols except 3 which are defined in libmagic.so and are exported by it. This works... Or at least libVersion() and LIB_VERSION values match (and it reports version 2 not 1).
-- However, when some data structures are serialized to disk, I noticed some corruption. In the application's directory if I delete libsome1.so and create a soft link in its place to point to libsome2.so, everything works as expected and the same corruption does not happen.
I can't help but think that this may be caused due to some conflict in the run-time linker's resolution of symbols. I've tried many things, like trying to link libsome2.so so that all symbols are alised to symbol##VER_2 (which I am still confused about because the command nm -CD libsome2.so still lists symbols as symbol and not symbol##VER_2)... Nothing seems to work!!! Help!!!!!!
Edit: I should have mentioned it earlier, but the app in question is Firefox, and libsome1.so is libsqlite3.so shipped with it. I don't quite have the option of recompiling them. Also, using version scripts to hide symbols seems to be the only solution right now. So what really happens when symbols are hidden? Do they become 'local' to the SO? Does rtld have no knowledge of their existence? What happens when an exported function refers to a hidden symbol?
Try compiling both libsome1.so and libsome2.so to add symbol versioning, each with their own version (use the --version-script option to ld). Then link the application and libmagic.so using the new libraries. Then, libsome1.so and libsome2.so should be completely separate.
Problems can still occur if there are unversioned references to symbols. Such references can be satisfied by versioned definitions (so that it is possible to add symbol versioning to a library without breaking binary compatibility). If there are multiple symbols of the same name, it can sometimes be hard to predict which one will be used.
Regarding tools, nm -D does not display any information about symbol versioning. Try objdump -T or readelf -s instead.

How to load all symbol-files recursively from given path including subdirectories?

You can point a single symbol file to gdb with command the:
symbol-file /usr/lib/debug/symbolfile.so
But how to tell gdb to load all symbol-files from given path including subdirectories?
On a Linux system, you should never have to use symbol-file GDB command in the first place.
The trick is to prepare your binaries in such a way that GDB will find the symbol file automatically. This is surprisingly easy to do. Detailed instructions are here.
Use following command:
set solib-search-path path
The solution is to add-symbol-file. For instance, if symbol file is called lib.out:
add-symbol-file lib.out 0
This is particularly useful on embedded system where application developers use a library stored in ROM. The debugger needs the symbol file to reconstruct the stack if execution stops in the middle of a library function call.
This works even if the library was generated on a separate system to which the developers have no access.

Resources