This is the mail archive of the
mailing list for the Archer project.
Re: Using Python to infer some internal data
2009/1/12 Daniel Jacobowitz <firstname.lastname@example.org>
> Tom will be able to give you a more definitive answer - but mine is
> that I don't think either is possible yet, but both of these are on
> my list of original goals for scripting support.
I thought as much, well I'll happily wait for Tom definite answer.
>>  http://bramp.net/blog/autoload-symbols-for-freebsd-kernel-module
> Caz, I've copied you because what Andrew's done here is very close to
> what you posted for Linux KGDB - but done entirely in Python
> scripting. Once we get the breakpoint issue resolved, I think this is
> a much nicer way to handle Linux loadable modules too.
I was unaware of Caz's work. Is that part of Kgdb Light? (A quick
Google just found that for me).
If you believe that my way is nicer, then I'll happily give you my
code to use license free.
2009/1/12 Caz Yokoyama <email@example.com>:
> I agree. It uses add-symbol-file to add symbols for loadable modules. I
> thought my code should use add-symbol-file (or underneath functions). My
> concern here is
> - Is there its counter part for Linux kernel?
I've only written code for FreeBSD, but I'm sure it be trivial to do
the same for Linux. I'm not familiar with the correct part of the
Linux kernel, but if you point me at which list/struct holds the names
of the loaded modules I can easily adapt my code.
> - Is the code which break-in kernel when gdb is starts valuable? Before gdb
> is started for Linux kernel, we have to type Ctrl+Alt+SysRq and g to break
> in to kernel. My code sends BREAK and g to kernel through serial port to
> break in. This code is mine and 16 lines.